Epidemiological and Clinical Profile of Patients Submitted to Total Knee Arthroplasty^[*]

Abstract

Objective To define the epidemiological profile and the associated comorbidities of patients submitted to total knee arthroplasty (TKA) in two reference hospitals.

Materials and Methods During the present cross-sectional observational study, 294 patients submitted to TKA were evaluated in 2 tertiary care hospitals. The diagnosis of self-reported comorbidities was collected by direct interview. The Functional Comorbidity Index (FCI) and the 5-Factor Modified Frailty Index (mFI-5) were calculated. The incidence of comorbidities and their relationship with the calculated indexes were evaluated.

Results Most of the patients in the sample were women (p = 0.000) between the seventh and eighth decades of life. Systemic arterial hypertension was the most common pathology, followed by obesity and diabetes mellitus. The FCI presented a direct relationship with females (p = 0.038) and obesity (p < 0.001). The mFI-5 was only associated to obesity (p = 0.022), demonstrating a higher chance of complications in this group.

Conclusion Patients undergoing TKA are essentially carriers of clinical comorbidities that can negatively influence functional results and compromise the safety of the procedure. The identification of risk factors can contribute to the safety and better selection of TKA candidates.

Keywords

arthroplasty - knee - comorbidity - epidemiology - postoperative complications

Introduction

Total knee arthroplasty (TKA) is the treatment option for patients with severe osteoarthritis (OA) refractory to conservative treatment. This surgery can improve pain by providing limb alignment and the return of mobility.[1] Most candidates for TKA are elderly patients with low physical demand and associated pathologies.[2]

Total knee arthroplasty is one of the most successful procedures of the last century.[3] More than 13 thousand procedures were performed in Brazil in 2019.[4] [5] This number should grow exponentially, driven by the aging of the population, increased obesity,[6] and the greater demand of the elderly population for quality of life.[7] Although safe, this surgery is considered large, and it is susceptible to possible complications.[8] A significant portion of patients remain dissatisfied after surgery. For these reasons, an index that can predict the functional outcome and the possible candidates with the greatest chance of complications will be extremely useful.

Comorbidities are defined as diseases or medical conditions with no causal relationship with the main diagnosis, coexisting with the pathology of interest.[9] Complications are adverse events that occur during a disease, and are not part of this pathology, although they may result from it.[10] The vast majority of patients who are candidates for TKA have associated clinical diseases that can significantly contribute to a higher rate of peri- and postoperative complications.[11] In order to predict the functional result after TKA and the possible chance of complication, respectively, two practical and objective questionnaires can be used: the Functional Comorbidity Index (FCI)[10] and the 5-Factor Modified Frailty index (mFI-5).[11]

Brazil has a large deficit in the care for arthroplasty surgeries, which are complex and costly procedures.[7] Anticipating the possible factors that lead to complications and worse functional results can contribute to a greater control of expenses and better patient selection. To this end, it is essential to know the profile of the patients submitted to TKA and the incidences of associated pathologies, improving the quality of care and avoiding unfavorable results.

The aim of the present study is to define the epidemiological profile of patients submitted to TKA and their associated comorbidities in two reference hospitals for this procedure.

Materials and Methods

The present cross-sectional observational study was approved by the Ethics in Research Committees of the institutions involved. All procedures were performed in accordance with the ethical standards of the 1964 Declaration of Helsinki , with their subsequent changes or comparable ethical standards. Patients were recruited from two high-complexity hospitals (tertiary care) that references for the TKA procedure. The sample consisted of patients from several socioeconomic levels. The ethnicity was diverse, as is characteristic of the Brazilian population. Although the phenotypic manifestation of skin color has little relation to ancestry, this characteristic was documented by self-report.

The Informed Consent Form (ICF) was signed by all patients before their inclusion in the study. Data were collected between August 2012 and December 2016. A total of 294 patients were prospectively evaluated, 203 female and 91 male. Data on comorbidities were collected by direct interview, with self-reported diagnosis. After the documentation, the FCI[10] and the mFI-5 were calculated.[11] As defined in the preoperative evaluation protocol, the patients were weighed and had their height measured wearing only underwear and an apron. After these measurements, body mass indexes were calculated, and a categorization was performed according to the parameters described by the World Health Organization (WHO). The radiographic OA degree was classified by the main author according to the criteria of the Ahlbäck classification modified by Keyes et al.[12]

The descriptive analysis of the data was based on graphs and on the calculation of descriptive statistics (proportions of interest, minimum, maximum, mean, median, and standard deviation). In the inferential analysis, statistical-significance tests were performed. Given the non-normality of the distributions of the variables FCI and mFI-5, a nonparametric approach was used. The Kruskall-Wallis nonparametric test was used to compare two independent groups, and the Mann-Whitney test was chosen for the comparison of more than two independent groups. The level of significance adopted was 5%. The association between two quantitative or ordinal variables was investigated by the Spearman order correlation coefficient. The significance of correlation coefficients was evaluated by the correlation coefficient test. The analyses were made using the Statistical Package for the Social Sciences (SPSS, IBM Corp., Armonk, NY, US) software, version 22.0, and the 2011 Microsoft Excel (Microsoft Corp., Redmond, WA, US) software.

Results

In total, 294 patients who met the inclusion criteria and agreed to participate in the study were submitted to TKA signing the ICF. The sample consisted of 203 (69%) female patients and 91 (31%) male patients, with a difference in the distribution in relation to gender (p = 0.000).

The age of the patients ranged from 40 to 86 years. The mean age was 69 years, 68.3 years for the male group and 69.4 years for the female group. There was no significant difference in the age group between the sexes (p = 0.239). Age statistics are shown in [Table 1].

Table 1

Age (years)	Female	Male	Global
Average	69.4	68.3	69.0
Median	70.0	68.0	69.0
Standard deviation	7.5	7.9	7.6
Minimum	40.0	43.0	40.0
Maximum	86.0	82.0	86.0
Range	46.0	39.0	46.0
Coefficient of variation	0.11	0.12	0.11
Number of cases	203	91	294
p-value of the Kolmogorov-Smirnov test	0.053	0.044	0.002
p-value of the Shapiro-Wilk test	0.032	0.054	0.0011
p-value of the Mann-Whitney test	0.293

Most patients in the sample declared themselves white (72.7%), followed by brown (15.9%) and black (11.4%). The body mass index (BMI) of the patients ranged from 17.3 kg/m² to 44.1 kg/m², with an average of 30 kg/m², standard deviation of 4.9 kg/m², median of 29.7 kg/m², and normal distribution in both groups. Women were significantly heavier than men (p = 0.017), although there was no significant difference in the proportion of obese individuals (p = 0.70). Analyses of the BMI of the global sample and of the sample divided by groups are shown in [Table 2].

Table 2

Weight Classification	Gender
	Female	Male	Global
Low weight	1 (0.5%)	0	1 (0.3%)
Normal weight	30 (15%)	18 (20.2%)	48 (16.6%)
Overweight	65 (32.5%)	35 (39.3%)	100 (34.6%)
Obesity I	64 (32.0%)	29 (32.6%)	93 (32.2%)
Obesity II	32 (16%)	6 (6.7%)	38 (13.1%)
Obesity III	8 (4.0%)	1 (1.1%)	9 (3.1%)

Comorbidities were analyzed by gender and their distribution in the total sample. There was no difference between the genders in the distribution of pathologies, except for rheumatoid arthritis, which was more frequent in the female group (p = 0.021). The main statistics of the distribution of pathologies are shown in [Table 3].

Table 3

Comorbidities	Incidence of comorbidity (%)			p-value of the test comparing the incidences in the female and male groups
	Female	Male	Global
Systemic arterial hypertension	67.5	58.2	64.6	0.125*
Obesity	52.0	40.4	48.4	0.070*
Diabetes mellitus	14.8	14.3	14.6	0.912*
Rheumatoid arthritis	5.9	0.0	4.1	0.021**
Dyslipidemia	4.4	2.2	3.7	0.512**
Cardiopathy	3.0	3.3	3.1	1.000**
Hypothyroidism	3.9	0.0	2.7	0.062**
Asthma	2.0	0.0	1.4	0.315**
Chronic obstructive pulmonary disease	1.0	1.1	1.0	1.000**
Glaucoma	0.5	2.2	1.0	0.227**
Venous insufficiency	1.5	0.0	1.0	0.555**

[Figure 1] illustrates the distribution of the most common pathologies in the total sample and their distribution by gender.

Fig. 1 Distribution of the most common pathologies in the total sample and their distribution by gender.

Most of the sample was aged between 60 and 70 years, and the comorbidities were also more frequent among this age group, although there was no statistically significant difference regarding the FCI (p = 0.221) and mFI-5 (p = 0.365) among the age groups. The distribution of pathologies by age group is shown in [Table 4].

Table 4

Age group/Comorbidities	41–50 years	51–60 years	61–70 years	71–80 years	> 80 years
Number	3	36	128	113	14
Systemic arterial hypertension	33.3%	56.6%	60.6%	69.9%	64.3%
Diabetes mellitus	–	16.7%	16.5%	20.4%	7.1%
Rheumatoid arthritis	33.3%	16.7%	0.8%	1.8%	7.1%
Dyslipidemia	–	2.8%	3.2%	4.4%	–
Cardiopathy	–	2.8%	3.1%	3.5%	–
Hypothyroidism	–	2.8%	3.1%	0.9%	7.1%
Asthma	–	0	0.8%	1.8%	–
Chronic obstructive pulmonary disease	–	2.8%	0.8%	0.9%	–
Glaucoma	–	–	–	2.7%	–
Venous insufficiency	–	–	0.8%	1.8%	–
Average Functional Comorbidity Index	1.7 ± 0.6	1.8 ± 0.6	1.8 ± 0.8	1.8 ± 0.8	1.6 ± 0.5
Median Functional Comorbidity Index	2.0	2.0	2.0	2.0	2.0
Average 5-Factor Modified Frailty Index	0.07 ± 0.12	0.16 ± 0.15	0.16 ± 0.14	0.19 ± 0.13	0.14 ± 0.12
Median 5-Factor Modified Frailty Index	0.00	0.20	0.20	0.20	0.20

The FCI presented a direct relationship with the female gender (p = 0.038) and with BMI (p< 0.001), but it was not related to age (p = 0.221), skin color (p = 0.058) or OA radiographic classification (Ahlbäck) (p = 0.420). The only variable that presented a significant correlation with the mFI-5 was BMI (p = 0.022), which was not related to gender (p = 0.237), age (p = 0.365), skin color (p = 0.251) or OA severity (p = 0.874).

Discussion

Total knee arthroplasty is one of the most performed surgeries in the world. In the United States, about 700 thousand TKAs are performed per year, reaching an average of 213.28 TKAs per 100 thousand inhabitants.[7] In Brazil, 13,210 TKAs were performed in 2019,[4] [5] representing 6.29 TKAs/100 thousand inhabitants. In addition to the more restricted indications and the different population profile, the lack of care and problems with the notification contribute to this large difference. There is an exponential increase, driven by population changes and greater access to health services, and it is important to direct resources to patients who will obtain greater benefits from the procedure.[13]

The sample consisted of a larger number of female patients, reflecting the incidence of OA, which affects two to three women for every man.[14] Although it is controversial in the literature, the satisfaction rate after TKA does not seem to diverge between men and women. Singh et al.,[15] after evaluating almost 18 thousand patients undergoing TA, found in men the highest chance of infection (+31%), readmission in the first 30 days (+25%), revision in the first 5 years (+20%), and the highest mortality rate in the first year (+48%). Robinson et al.[16] studied 32,848 patients undergoing TKA, and concluded that female gender is a protective factor against sepsis and cardiovascular and oral complications, although it presents a higher risk of developing urinary infection, of needing blood transfusion, and of presenting difficulty supporting loads on the operated limb.

For the evaluation of the surgical results, it is important to know the epidemiological profile of the treated patients, because several clinical comorbidities can directly impact the outcome,[17] compromising metabolic demand, wound healing, and anesthesia administration.[18] Despite the benefits associated with TKA, up to 24% of the operated patients may face serious adverse events, and it is essential to consider who should undergo the procedure. The preoperative evaluation aims to clinically optimize the patients, minimizing their risks and facilitating their recovery.[19]

We used self-reported diagnoses as a tool to list comorbidities because they are an attractive method for chronic diseases. Najafi et al.[20] described this strategy as reliable and efficient, presenting some advantages over the analysis of the medical records, as it does not depend on the bias of the notifier. Furthermore, for administrative and contractual reasons, these pathologies may be omitted and underreported. All patients in our sample were treated by the Brazilian Unified Health System (Sistema Único de Saúde, SUS, in Portuguese), which is ideal for this type of study, because they did not suffer pressure from the insurer, and there was no influence regarding the omission of pathologies.

The most frequently found comorbidity was systemic arterial hypertension (SAH), which affected 64.9% of the sample. The highest incidence occurred in the eighth decade of life (69.9%). The prevalence of the disease was similar the in male and female groups (p = 0.125). This result is consistent with that described by Feng et al.,[21] who highlighted that SAH is the most common comorbidity in the orthopedic practice. The literature is scarce in studies on the possible influence of SAH on the outcome of patients undergoing TKA. Elmallah et al.[22] found worse functional results in hypertensive patients when compared to normotensive patients, although the improvement in pain was similar. Thus, SAH can negatively influence the functional outcome of surgery. Dyslipidemia affected less than 5% of the patients, being also more frequent in the eighth decade of life. The use of statins should be continued, as there is evidence that this medication may cause a decrease both in the early mortality rate and in the chance of aseptic loosening of the prosthesis in the first ten years.[19]

Diabetes mellitus (DM) was present in 14.6% of the total sample, being higher in the eighth decade of life (20.4%). There was no difference between gender groups (p = 0.912). According to Flor and Campos,[23] the incidence in the Brazilian population older than 65 years of age is of 16.7%. The sample included patients below this age group, between 41 and 50 years old, who did not present DM and influenced the total mean. Diabetes mellitus is a pathology that interferes with wound healing, decreases osteoblastic activity, and impairs the immune system.[19] Diabetic patients have an increase in the rate of complications, such as periprosthetic infection (relative risk, RR: 1.6), deep venous thrombosis (RR: 2.57) and aseptic loosening (RR: 9.36).[24] Singh and Lewallen[25] evaluated 7,139 diabetic patients submitted to TKA, and concluded that DM is an independent risk factor for unfavorable results after surgery. Despite the higher number of complications and poorer functional outcomes, diabetic patients have a satisfaction rate similar to that of non-diabetic patients, and the procedure is able to generate other benefits, such as a reduction in BMI of 50% among those operated.[26] Attention should be given to preoperative glycemic levels, with the value of 200 mg/dL being the limit.[27]

The analysis of age revealed that 82% of the patients are between the seventh and eighth decades of life. These data reflect the stricter indications for this surgery in Latin patients. Although younger patients have a faster recovery and excellent functional results,[28] they have a lower satisfaction rate and a higher risk of revision. Life expectancy in Brazil has increased exponentially, reaching 76.74 years for births in 2020. Projections indicate that by 2050 it will reach 81.29 years. Cher et al.[29] evaluated 209 octogenarians submitted to TKA and found excellent functional results, not differing from those of their younger peers. However, the authors point out that, alongside the improvement in quality of life, there is an increase in the cost to offer surgical treatment to increasingly elderly patients. Each year of age added increases the chance of rehospitalization and mortality.[30]

Obesity is a factor that is clearly involved in the genesis and progression of knee OA.[6] This fact was proven in our sample, in which only 16.9% of the patients were normotrophic, with the remainder being overweight or obese. Female patients were associated with higher BMIs (p = 0.017). Obesity is an independent risk factor for postoperative and perioperative complications, such as acute myocardial infarction, stroke, periprosthetic infection and need for revision,[19] [21] in addition to the greater chance of rehospitalization and mortality.[30] Several institutions have restricted TKA in patients with BMI greater than 40 kg/m² due to the higher cost and chance of unfavorable events.[21] Weight control should be encouraged for patients with BMI greater than 25 kg/m², and bariatric surgery may be a pre-TKA resource for obese patients refractory to the conservative treatment. Traven et al.[11] created a protocol with five factors (mFI-5) capable of foreseeing the chance of complication after TKA. In the sample studied, obesity was the only factor directly related to mFI-5 (p = 0.022), proving the relationship of excess body weight and a worse prognosis.

Despite the success of TKA in restoring function and improving quality of life, up to 20% of patients may experience postoperative pain or dissatisfaction. Several authors have developed tools to predict the functional result. Groll et al.[10] developed the FCI. Unlike other indexes, which assess the time of hospitalization and the mortality rate, the FCI indicates the probable functional outcome of the patients, with 77% of efficacy. The authors included 18 pathologies in the questionnaire, which may be self-applicable. The questions include the evaluation of mental health and respiratory diseases, which are important factors related to the outcome of the TKA. The sample studied presented a worse functional prognosis for obese individuals (p < 0.001) and female patients (p = 0.038), which is consistent with the findings by Robinson et al.,[16] who described a greater difficulty of ambulation after surgery in female and obese patients.[16]

The present study has some limitations. Because it is a cross-sectional observational study, the patients did not have their outcome analyzed. However, we believe that the results meet the objective of defining the epidemiological profile of patients undergoing TKA. The instruments included, the FCI and the mFI-5, are interesting tools, but they are not able to establish a limit after which TKA should not be performed. Therefore, the participation of patients and family members in the decision-making process should be encouraged.

Conclusion

Patients undergoing TKA are essentially carriers of clinical comorbidities that can negatively influence the functional outcomes and increase the rate of surgery complications. The identification of risk factors contributes to the safety and better selection of candidates.

Conflito de Interesses

Os autores declaram não haver conflito de interesses.

Financial Support

There was no financial support from public, commercial, or non-profit sources.

^* Study developed at the Knee Surgery Group, Hospital Santa Teresa, Petrópolis, RJ, Brazil.

• Referências

• 1 da Silva RR, Santos AA, de Sampaio Carvalho Júnior J, Matos MA. Quality of life after total knee arthroplasty: systematic review. Rev Bras Ortop 2014; 49 (05) 520-527
• 2 Zanasi S. Innovations in total knee replacement: new trends in operative treatment and changes in peri-operative management. Eur Orthop Traumatol 2011; 2 (1-2): 21-31
• 3 Schwechter EM, Fitz W. Design rationale for customized TKA: a new idea or revisiting the past?. Curr Rev Musculoskelet Med 2012; 5 (04) 303-308
• 4 Ministério da Saúde. Secretaria Executiva. Datasus [acesso em: 28 de março de 2020]. Informações de Saúde; . Disponível em: http://tabnet.datasus.gov.br/cgi/tabcgi.exe?sih/cnv/qiuf.def
• 5 Agência Nacional de Saúde Suplementar ANS [acesso em 28 de Março de 2020]. Painel de dados do TISS. 2018 . Disponível em: https://app.powerbi.com/view?r=eyJrIjoiMDNlMDNlNzEtZDA5Yy00YzE5LTg2Y2ItY2IxNTBkMTIwM2Q5IiwidCI6IjlkYmE0ODBjLTRmYTctNDJmNC1iYmEzLTBmYjEzNzVmYmU1ZiJ9
• 6 Loures FB, de Araújo Góes RF, Labronici PJ, Barretto JM, Olej B. Evaluation of body mass index as a prognostic factor in osteoarthrosis of the knee. Rev Bras Ortop 2016; 51 (04) 400-404
• 7 Ferreira MC, Oliveira JCP, Zidan FF, Franciozi CES, Luzo MVM, Abdalla RJ. Artroplastia total de joelho e quadril: a preocupante realidade assistencial do Sistema Único de Saúde brasileiro. Rev Bras Ortop 2018; 53 (04) 432-440
• 8 Bjorgul K, Novicoff WM, Saleh KJ. Evaluating comorbidities in total hip and knee arthroplasty: available instruments. J Orthop Traumatol 2010; 11 (04) 203-209
• 9 Healy WL, Della Valle CJ, Iorio R. et al. Complications of total knee arthroplasty: standardized list and definitions of the Knee Society. Clin Orthop Relat Res 2013; 471 (01) 215-220
• 10 Groll DL, To T, Bombardier C, Wright JG. The development of a comorbidity index with physical function as the outcome. J Clin Epidemiol 2005; 58 (06) 595-602
• 11 Traven SA, Reeves RA, Sekar MG, Slone HS, Walton ZJ. New 5-Factor Modified Frailty Index Predicts Morbidity and Mortality in Primary Hip and Knee Arthroplasty. J Arthroplasty 2019; 34 (01) 140-144
• 12 Keyes GW, Carr AJ, Miller RK, Goodfellow JW. The radiographic classification of medial gonarthrosis. Correlation with operation methods in 200 knees. Acta Orthop Scand 1992; 63 (05) 497-501
• 13 Dakin H, Gray A, Fitzpatrick R, Maclennan G, Murray D. KAT Trial Group. Rationing of total knee replacement: a cost-effectiveness analysis on a large trial data set. BMJ Open 2012; 2 (01) e000332
• 14 Ro DH, Lee DY, Moon G. et al. Sex differences in knee joint loading: Cross-sectional study in geriatric population. J Orthop Res 2017; 35 (06) 1283-1289
• 15 Singh JA, Kwoh CK, Richardson D, Chen W, Ibrahim SA. Sex and surgical outcomes and mortality after primary total knee arthroplasty: a risk-adjusted analysis. Arthritis Care Res (Hoboken) 2013; 65 (07) 1095-1102
• 16 Robinson J, Shin JI, Dowdell JE, Moucha CS, Chen DD. Impact of Gender on 30-Day Complications After Primary Total Joint Arthroplasty. J Arthroplasty 2017; 32 (08) 2370-2374
• 17 Santos AC, Biagi AC. Epidemiological profile of patients undergoing knee arthroplasty from the physiotherapy service of a public hospital in Santo André-SP. ABCS Health Sci 2013; 38 (01) 2-7
• 18 Skou ST, Roos EM, Laursen MB. et al. A Randomized, Controlled Trial of Total Knee Replacement. N Engl J Med 2015; 373 (17) 1597-1606
• 19 Adie S, Harris I, Chuan A, Lewis P, Naylor JM. Selecting and optimising patients for total knee arthroplasty. Med J Aust 2019; 210 (03) 135-141
• 20 Najafi F, Moradinazar M, Hamzeh B, Rezaeian S. The reliability of self-reporting chronic diseases: how reliable is the result of population-based cohort studies. J Prev Med Hyg 2019; 60 (04) E349-E353
• 21 Feng JE, Novikov D, Anoushiravani AA, Schwarzkopf R. Total knee arthroplasty: improving outcomes with a multidisciplinary approach. J Multidiscip Healthc 2018; 11: 63-73
• 22 Elmallah RD, Cherian JJ, Robinson K, Harwin SF, Mont MA. The Effect of Comorbidities on Outcomes following Total Knee Arthroplasty. J Knee Surg 2015; 28 (05) 411-416
• 23 Flor LS, Campos MR. Prevalência de diabetes mellitus e fatores associados na população adulta brasileira: evidências de um inquérito de base populacional. Rev Bras Epidemiol 2017; 20 (01) 16-29
• 24 Yang Z, Liu H, Xie X, Tan Z, Qin T, Kang P. The influence of diabetes mellitus on the post-operative outcome of elective primary total knee replacement: a systematic review and meta-analysis. Bone Joint J 2014; 96-B (12) 1637-1643
• 25 Singh JA, Lewallen DG. Diabetes: a risk factor for poor functional outcome after total knee arthroplasty. PLoS One 2013; 8 (11) e78991
• 26 Teo BJX, Chong HC, Yeo W, Tan AHC. The Impact of Diabetes on Patient Outcomes After Total Knee Arthroplasty in an Asian Population. J Arthroplasty 2018; 33 (10) 3186-3189
• 27 Parvizi J, Shohat N, Gehrke T. Prevention of periprosthetic joint infection: new guidelines. Bone Joint J 2017; 99-B (4, Supple B) 3-10
• 28 Kleeblad LJ, van der List JP, Zuiderbaan HA, Pearle AD. Larger range of motion and increased return to activity, but higher revision rates following unicompartmental versus total knee arthroplasty in patients under 65: a systematic review. Knee Surg Sports Traumatol Arthrosc 2018; 26 (06) 1811-1822
• 29 Cher EWL, Tay KS, Zhang K, Tan SB, Howe TS, Koh JSB. The Effect of Comorbidities and Age on Functional Outcomes After Total Knee Arthroplasty in the Octogenarian: A Matched Cohort Study. Geriatr Orthop Surg Rehabil 2018; 9: 2151459318769508
• 30 Lehtonen EJ, Hess MC, McGwin Jr G, Shah A, Godoy-Santos AL, Naranje S. Risk factors for early hospital readmission following total knee arthroplasty. Acta Ortop Bras 2018; 26 (05) 309-313

Publication History

Received: 03 June 2020

Accepted: 02 October 2020

Article published online:
31 March 2021