Physical Activity and Sedentary Behaviour of Adults with Type 2 Diabetes: a Systematic Review

Review

Physical activity and sedentary behaviour of adults with type 2 diabetes: a systematic review

Ann-Marie Kennerly1 Abstract BSc (Hons), Student at University of Strathclyde Regular physical activity and low levels of prolonged sedentary behaviour are important lifestyle recommendations for the management of type 2 diabetes. This systematic review 1 Alison Kirk collates and summarises published research reporting on physical activity and sedentary BSc (Hons), PhD, Senior Lecturer in Physical Activity behaviour of adults with type 2 diabetes. for Health Systematic searches with three databases were conducted (Medline; PubMed; 1School of Psychological Sciences and Health, SportDiscus). Intervention studies were excluded. Studies were eligible if they: (i) were University of Strathclyde, Glasgow, UK published in English; (ii) were published between 1997 and January 2017; and (iii) objectively and/or subjectively measured physical activity and/or sedentary behaviour in adults with type 2 diabetes. Two reviewers independently cross-checked studies, conducted a Correspondence to: quality assessment and extracted data. Data were analysed using descriptive statistics and a Dr Alison Kirk, Room 532 Graham Hills Building, narrative synthesis. School of Psychological Sciences and Health, The search identified 349 studies; 29 eligible studies were included. All studies measured University of Strathclyde, Glasgow G1 1QX, UK; physical activity and 15 studies measured sedentary behaviour. Twenty studies used email: [email protected] subjective methods, six used objective methods and three used both subjective and objective methods. Most studies report low levels of physical activity and high levels of sedentary Received: 24 November 2017 Accepted in revised form: 13 March 2018 behaviour in adults with type 2 diabetes and note adults with type 2 diabetes to be less active and more sedentary than those without type 2 diabetes. Regardless of measurement method or study location, adults with type 2 diabetes report low levels of physical activity and high levels of sedentary behaviour and are less active and more sedentary than those without type 2 diabetes. There is a need for large-scale interventions which support active lifestyles to be implemented into diabetes care. Copyright © 2018 John Wiley & Sons. Practical Diabetes 2018; 35(3): 86–89 Key words physical activity; sedentary behaviour; type 2 diabetes

Background to meet current PA guidelines of at There is substantial evidence docu- least 150 minutes a week of moderate menting the wide-ranging benefits PA or 75 minutes of vigorous intensity of frequent, regular physical activity exercise per week.6 Reducing the (PA) for management of type 2 dia- amount of time spent sedentary and betes (T2D) with improvements in breaking up prolonged SB of >30 min- glycaemic control, cardiovascular utes with light activity is also recom- risk factors and overall quality of life mended for management of T2D.1 being reported.1 Independent of PA Surveys using self-report meth- levels, time spent in sedentary ods have found adults with T2D are behaviour (SB) has also been shown less physically active than those with- to be deleterious for health.2 out the condition.7 Recently, data Sedentary behaviour is character- exploring this topic have been ised as any waking behaviour in a reported using objective methods. sitting or reclined posture with Both methods have advantages and energy expenditure of less than disadvantages. Objective methods 1.5 metabolic equivalents.3 A ran- can capture detailed information on domised crossover trial by Dempsey the pattern, for example intensity et al.4,5 concluded that interrupting and time of behaviour; however, prolonged SB with 3 minutes of they do not provide information light-intensity PA improves glucose about the type of activity and may metabolism in adults with T2D. not capture all physical activity such The American Diabetes Association as water-based activity. Subjective recommends that adults with T2D methods provide information on should exercise daily or at least every the type of activity but are often other day taking part in both aerobic criticised on issues related to robust- and anaerobic activities;1 the aim being ness of self-report data.

The objective of this systematic review was to review published evidence exploring PA and SB levels Records identified from in adults with T2D. database search (n=415): Additional studies identified PubMed (n=77) through other sources (n=0) Methods SportDiscus (n=198) The protocol for this systematic Medline (n=140) review can be accessed via www.crd. york.ac.uk registration number: CRD42017053717. A literature search was conducted in January Studies after duplicates 2017 to identify eligible studies. A removed (n=349) date restriction of 1997 was used to ensure findings were relevant to current practice. Only peer Studies removed after application Studies for title/abstract reviewed, published articles in of exclusion/inclusion criteria screening (n=349) English were included. Three data- (n=310) bases were searched using a Boolean search strategy: Medline, PubMed and SportDiscus. Search terms were: No full study available (n=3) Full-text articles assessed for • Population – over 18 years of age Intervention study (n=4) eligibility (n=39) with T2D. Participant selection bias (n=1) • Intervention – no intervention. • Comparison – no comparison or comparison with population without T2D or another relevant chronic Final studies eligible for Intervention studies (n=2) disease. qualitative synthesis (n=31) • Outcomes – SB or PA/inactivity as an outcome. • Study design – quantitative studies including cross-sectional Final studies included in and cohort design. systematic review (n=29)

A Medical Subject Headings (MeSH) search of T2D was con- Figure 1. PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) flow ducted to inform search terms. diagram illustrating the exclusion process and the final study selection into systematic review Definitions of PA and SB were used along with detailed guidance. measured, defined and analysed. to describe outcomes – for example: Differences in quality assessment Therefore, data were synthesised exercise; exercise habits; physical were resolved by discussion. Each and presented narratively. inactivity; sedentary time; sitting. question in the assessment tool was Additionally, reference lists and first rated either yes, no or other (can- Results authors of eligible studies were man- not determine [CD], not applicable Study selection ually searched. Intervention studies [NA], not reported [NR]). One Figure 1 illustrates the process of were excluded as PA behaviour point was available for each ques- article selection, including reasons could be non-representative of tion resulting in a possible 14-point for exclusion in accordance with participants’ usual patterns due to total for a high-quality study. Studies the Preferred Reporting Items for motivations during interventions. which scored below six points were Systematic reviews and Meta- Exclusion and inclusion criteria regarded to be of ‘poor’ quality, Analyses (PRISMA) statement. were applied to the titles and studies between 6–13 points were abstracts of studies and 100% of rated ‘fair’ and studies which scored Study characteristics the studies were cross-checked by 14 points were considered a ‘good’ The study characteristics and quality an independent researcher (AK). quality study. assessment are summarised in Disagreements were discussed and Alongside quality assessment, Appendix 1 (available online at www. agreed by consensus. relevant data were extracted. The practicaldiabetes.com) and are The quality of studies was extracted information included: grouped by methods of PA/SB meas- assessed individually based on year and location of publication; urement. The countries of included details reported by two reviewers sample characteristics and compar- studies were: America (n=8), Nigeria (AMK and AK) using the National ison group; measurement method; (n=3), the UK (n=3), Australia (n=2), Institute for Health assessment tool PA and/or SB results; and relevant Canada (n=2), Belgium (n=1), Brazil for observational, cohort and conclusions. There was incon (n=1), Croatia (n=1), Denmark cross-sectional studies. This assess- sistency across eligible studies (n=1), Finland (n=1), France (n=1), ment tool consists of 14 questions regarding how PA and SB were Nepal (n=1), Netherlands (n=1),

Poland (n=1), South Africa (n=1) age appropriate comparison group 15–61% from studies using the and Turkey (n=1). of people without T2D report IPAQ with IPAQ scoring protocol Three studies used both subjec- lower PA levels using a variety of and between 25–75% for other tive and objective methods. None of subjective methods in people with self-report measures. Regardless of these studies included a comparison T2D.7,17–22 Two of the three stud- the measurement method, adults sample.8–10 Only one study meas- ies23,24 which report a comparison with diabetes report lower levels of ured subjective and objective PA with people with type 1 diabetes PA compared to adults without levels over the same time period.10 report people with T2D to have diabetes. Physical activity is higher Six studies used objective methods lower levels of PA; however, people during weekdays compared to with five of these studies including a with type 1 diabetes are consistently weekend days and in spring/sum- comparison with an age appropriate younger across all three studies. mer compared to autumn/winter. non-diabetic sample.11–16 Two objec- One study which tracked changes tive methods of measurement were Sedentary behaviour. The four in PA across five years reported no used to assess PA and SB: accelerom- studies which objectively measured change in adults with diabetes over eters and pedometers. Twenty stud- sedentary time report variable out- time, in comparison to a significant ies used solely subjective methods comes including: minutes per day increase in people without dia across a variety of subjective instru- sedentary (659.6–926 minutes);11,16 betes. In comparison to adults with- ments to record participants’ activ- and proportion of waking day spent out diabetes, those with diabetes ity data.7,17–35 The International sedentary (64.5–75.1%).13,15,16 All have higher levels of SB, reporting Physical Activity Questionnaire three studies which included a around 5.5 hours of SB per day and (IPAQ) was the most common comparison group report people with approximately 70% of the waking resource.17,18,23,26–33 Four studies T2D to have higher levels of SB than day in sedentary behaviours. used researcher-developed ques- those without T2D.11,13,15 Studies Cassidy et al.17 explored self- tionnaires.19,25,34,35 Seven studies which subjectively measured SB reported PA and SB across cardio- using subjective methods included a report variable outcomes including: metabolic disease using a large popu- comparison with an age appropriate >3 hours/day watching TV (38.1%);17 lation-based cohort (UK Biobank). non-diabetic group7,17–22 and three mean sitting time/day (4.8–5.8 Four groups were defined: no dis- with a sample of people with type 1 hours);27,33 and mean sitting time/ ease (103 993); cardiovascular dis- diabetes.23–25 All 29 studies included week (302 minutes/week).26 ease (113 469); T2D (4074); CVD + in the systematic review measured T2D (11 574). Total PA declined PA levels, with eight studies also Discussion across these groups and total SB (TV recording SB.11,13,15–17,26,27,33 Four This study systematically reviews viewing) increased. The odds ratio of of the eight studies11,13,15,16 that published research documenting reporting low moderate PA (mins/ measured SB used objective meth- PA and SB in adults with T2D. day) was 1.18 (CVD), 1.35 (T2D) ods. Six studies received a ‘good’ Results highlight that, regardless and 1.55 (CVD + T2D), and the odds quality assessment rating.8–10,13,15,16 of measurement method, data ratio of reporting high SB was 1.42 For most studies (23/29) the quality reported or study location, adults (CVD), 1.59 (T2D) and 1.92 (CVD + assessment rating was either fair or with T2D report low levels of PA and T2D). The study demonstrates that poor (fair n=12, poor n=11). high levels of SB and are less active those with more advanced cardio- and more sedentary than those with- metabolic disease undertake too Summary of study findings out T2D. It is likely that multiple little PA and have high TV viewing Physical activity. Studies which factors play a role in influencing PA times, yet report important positive objectively measure PA report varia- and SB in people with T2D includ- dietary changes within the past five ble outcomes including: average ing chronic physical and mental years. These behaviours are clus- daily step count (range 4901–5659 health complications and age, in tered, highlighting the importance steps/day);8,9 Actigraph accelerom- addition to societal influences. of interventions focused on chang- eter count (140 302 to 155 520);14 Average step count was around ing multiple lifestyle behaviours. minutes of moderate to vigorous PA 5000 steps/day and time spent step- Although 29 studies were eligible (range 24.2–34 minutes/day);9,11 ping occupied approximately 10% for inclusion in this systematic review percentage meeting 7100 steps/day of the waking day. Accelerometer- from a wide range of locations, only (29.7%);12 percentage of waking measured minutes of moderate to three studies were from the UK. day spent stepping (10.7%).15 All vigorous PA was around 30 min- Overall, the quality of most existing five studies which used objective utes/day, occupying a small pro- studies is low. Studies use different methods of measurement and portion (0.8%) of the waking day. methods to process and report included a comparison group The percentage of adults with dia- data across different recording report people with T2D to be less betes meeting current PA recom- time periods making comparison active than age comparable people mendations of at least 150 minutes across studies difficult. Only three without diabetes.11–15 Studies which of moderate to vigorous PA per studies used objective and subjective subjectively measured PA report week varies, with objective data methods of PA measurement; variable outcomes even when using reporting approximately 9% meet- however, none of the eligible studies the same method of measurement. ing these recommendations while used both subjective and objective All seven studies which included an self-reported data vary between methods to measure PA and SB

patterns simultaneously. Some stand- 20. Zhao G, et al. Compliance with physical activity ardisation of data collection meth- Key points recommendations in US adults with diabetes. Diabet Med 2008;25:221–7. ods and reporting is required to 21. Zhao G, et al. Physical activity in U.S. older adults allow more accurate global compari- ● This study systematically reviews with diabetes mellitus: prevalence and correlates of sons to be conducted. published research reporting on meeting physical activity recommendations. J Am Overall, there is a need for large- physical activity and sedentary Geriatr Soc 2011;59:132–7. 22. Babić Z, et al. Estimation of physical activity by dif- scale interventions to be imple- behaviour in adults with type 2 ferent questionnaires in overweight subjects and mented into diabetes care which diabetes patients with type 2 diabetes mellitus: relationship support people with T2D to initiate, ● The results highlight that regardless of with anthropometric and metabolic variables. maintain and achieve the substantial measurement method, data reported or Diabetes Nutr Metab 2004;17:280–9. 23. Duarte CK, et al. Physical activity level and exercise benefits of an active lifestyle. study location, adults with type 2 in patients with diabetes mellitus. Rev Assoc Med diabetes have low levels of physical Bras (1992) 2012;58:215–21. Declaration of interests activity and high levels of sedentary 24. Plotnikoff RC, et al. Factors associated with physical activity in Canadian adults with diabetes. Med There are no conflicts of interest behaviour and are less active and Sci Sports Exerc 2006;38:1526–34. declared. more sedentary than those without 25. Thomas N, et al. Barriers to physical activity in type 2 diabetes patients with diabetes. Postgrad Med J 2004;80: References 287–91. 1. Colberg SR, et al. Physical activity/exercise and 26. Çolak TK, et al. Association between the physical diabetes: A position statement of the American 11. Cichosz SL, et al. Objective measurements of activ- activity level and the quality of life of patients Diabetes Association. Diabetes Care 2016;39: ity patterns in people with newly diagnosed type 2 with type 2 diabetes mellitus. J Phys Ther Sci 2065–79. diabetes demonstrate a sedentary lifestyle. Diabet 2016;28:142–7. 2. Wilmot EG, et al. Sedentary time in adults and the Med 2013;30:1063–6. 27. Cooper J, et al. Self-reported physical activity in association with diabetes, cardiovascular disease 12. Kelly J, et al. Gender differences in physical activity medically underserved adults with type 2 diabetes and death: systematic review and meta-analysis. levels of older people with type 2 diabetes mellitus. in clinical and community settings. J Phys Act Diabetologia 2012;55:2895–905. J Physical Activity & Health 2016;13:409–15. Health 2015;12:968–75. 3. Sedentary Behaviour Research Network. October 2, 13. Loprinzi PD. Accelerometer-determined sedentary 28. Nolan RC, et al. Self-reported physical activity using 2015. Available from: /www.sedentarybehaviour. and physical activity estimates among older adults the International Physical Activity Questionnaire org. with diabetes: Considerations by demographic and (IPAQ) in Australian adults with type 2 diabetes, 4. Dempsey PC, et al. Benefits for type 2 diabetes comorbidity characteristics. J Aging Phys Act 2014; with and without peripheral neuropathy. Can J of interrupting prolonged sitting with brief bouts 22:432–40. Diabetes 2016;40:576–9. of light walking or simple resistance activities. 14. Steeves JA, et al. Daily patterns of physical activity 29. Oguntibeju OO, et al. Health behaviour and quality Diabetes Care 2016;39:964–72. by type 2 diabetes definition: Comparing diabetes, of life of patients with type 2 diabetes attending 5. Dempsey PC, et al. Interrupting prolonged sitting in prediabetes, and participants with normal glucose selected hospitals in south western Nigeria. West type 2 diabetes: Nocturnal persistence of improved levels in NHANES 2003–2006. Prev Med Rep 2015; Indian Med J 2012;61:619–26. glycaemic control. Diabetologia 2016;39:964–72. 2:152–7. 30. Ojoawo AO, Ogunbeku MA. Assessment of level of 6. Department of Health. Start Active, Stay Active: A 15. Van der Berg JD, et al. Associations of total amount physical activity in patients with type 2 diabetes report on physical activity from the four home and patterns of sedentary behaviour with type 2 mellitus. Sports Med J / Medicina Sportivâ 2012;8: countries’ Chief Medical Officers. London: diabetes and the metabolic syndrome: The 1750–5. Department of Health, 2011. Maastricht Study. Diabetologia 2016;59:709–18. 31. Cloix L, et al. Physical activity at home, at leisure, 7. Morrato EH, et al. Physical activity in U.S. adults 16. Hamer M, et al. Objectively assessed physical activ- during transportation and at work in French adults with diabetes and at risk for developing diabetes, ity, adiposity, and inflammatory markers in people with type 2 diabetes: the ENTRED physical activity 2003. Diabetes Care 2007;30:203–9. with type 2 diabetes. BMJ Open Diabetes Res & study. Diabetes Metab 2015;41:37–44. 8. Dasgupta K, et al. Daily steps are low year-round and Care 2014;2:1. 32. Kabanda AM, Phillips JS. Physical activity among dip lower in fall/winter: findings from a longitudinal 17. Cassidy S, et al. Cross-sectional study of diet, phys- adults with diabetes mellitus in Rwanda. African J diabetes cohort. Cardiovasc Diabetol 2010;9:81. ical activity, television viewing and sleep duration Physical, Health Education, Recreation & Dance 9. De Greef K, et al. Physical environmental correlates in 233,110 adults from the UK Biobank; the behav- 2011;17:239–47. of self-reported and objectively assessed physical ioural phenotype of cardiovascular disease and 33. Oyewole OO, et al. Physical activity among type-2 activity in Belgian type 2 diabetes patients. Health type 2 diabetes. BMJ Open 2016;6:e010038. diabetic adult Nigerians. Ann Afr Med 2014;13: Soc Care Community 2011;19:178–88. 18. Sibai AM, et al. Physical activity in adults with and 189–94. 10. Mynarski W, et al. Declared and real physical activ- without diabetes: from the ‘high-risk’ approach to 34. Hays LM, Clark DO. Correlates of physical activity in ity in patients with type 2 diabetes mellitus as the ‘population-based’ approach of prevention. a sample of older adults with type 2 diabetes. assessed by the International Physical Activity BMC Public Health 2013;13:1002. Diabetes Care 1999;22:706–12. Questionnaire and Caltrac accelerometer monitor: 19. Karjalainen J, et al. Leisure time physical activity in 35. Parajuli J, et al. Factors associated with nonadher- a potential tool for physical activity assessment in individuals with screen-detected type 2 diabetes ence to diet and physical activity among Nepalese patients with type 2 diabetes mellitus. Diabetes Res compared to those with known type 2 diabetes. type 2 diabetes patients; a cross sectional study. Clin Pract 2012;98:46–50. Diabetes Res Clin Pract 2008;81:110–6. BMC Res Notes 2014;7:758.

First author, Country Type 2 Comparison Method of Results: mean/median (SD) Relevant Quality year diabetes sample measurement conclusions assessment sample rating†

No. + mean No. + mean Physical Sedentary age (SD) age (SD) activity behaviour

Combined objective and subjective method of PA/SB measurement

Dasgupta Canada T2D (n=166). – Yamax SW-701 Average daily – Daily step Good (2010)8 Mean (SD) pedometer + step count: counts in T2D age 62.4 IPAQ-SF (IPAQ Spring/summer patients are (10.8) scoring protocol 5659 (2611); low, dipping used) Fall/winter lower during 4901 (2464); fall/winter SR MET-mins/ week 1965

De Greef Belgium T2D (n=133). – IPAQ-LF. IPAQ MVPA – – Good (2011)9 Mean (SD) 80.9 (158.2) age Actigraph 7164 min/wk. 61.6 (8.4) accelerometer. AC MVPA 24.2 (28.3) Yamax DigiWalker min/day. SW200 Step counts pedometer 5365 (3070) steps/day

Mynarski Poland T2D (n=31). – IPAQ-SF. IPAQ PA – No significant Good (2012)10 Mean (SD) 2469±1326 difference age Caltrac MET-min/wk. between 54 (3.6) accelerometer CPA self-reported 4284±2146 and objective cal/wk measured PA

Objective method of PA/SB measurement

Cichosz Denmark T2D (n=94). ND (n=85). Actiheart MVPA: Min/day spent People with T2D Fair (2013)11 Mean (SD) Mean (SD) accelerometer. ND 62min/day; sedentary: significantly age 58 (10) age 57 (10) Accelerometer T2D 34min/day ND 926min/day less PA and years years count T2D 898min/ more SB than classification: day ND Sedentary <20cpm MVPA >20cpm

Kelly Australia T2D (n=293). ND (n=336). Yamax Digiwalker PA guideline of – In women, but Fair (2016)12 Mean age Mean age SW-200 7100 steps per not men, T2D (SD) 67.6 (6.8) (SD) pedometer day was met by: patients take 72.1 (7.1) T2D 29.7% fewer steps/ ND 33.3% day and less likely to meet current PA guidelines than ND

Legend: SD = standard deviation; T2D = type 2 diabetes; ND = no diabetes; D = diabetes; MVPA = moderate to vigorous PA; PA = physical activity; IPAQ-SF = International PA Questionnaire-short form; IPAQ-LF = International PA Questionnaire-long form; LRCPA = lipid research clinics PA; CPM = counts per minute; CPA = Caltrac PA; SR = self-report; AC = accelerometer based; h/wk= hours per week; min/day = minutes per day; min/wk = minutes per week; MET= metabolic equivalents; MET-min/wk = metabolic equivalent minutes per week; kcal/wk = calories per week; GLTEQ = Godin Leisure Time Exercise Questionnaire; TAC = total activity counts; BRFSS = Behavioural Risk Factor Surveillance System; ADA = American Diabetes Association.

†The National Institute for Health tool for observational, cohort and cross-sectional studies was used. *IPAQ scoring protocol: Low – failure to meet criteria for the ‘moderate’ or ‘high’ categories; Moderate – doing at least 150min/week of moderate-to-vigorous PA over at least 3 days; High – vigorous-intensity activity on at least 3 days and accumulating at least 1500 MET-min/week, or 7 or more days of any combination of walking moderate-intensity or vigorous-intensity activities, achieving a minimum of 3000 MET-min/week. ** Sport Index ranging from 1.0 = lowest PA to 5.0 = highest possible PA.

Appendix 1. Study characteristics and quality assessment of included research. (Continued on the next 6 pages)

First author, Country Type 2 Comparison Method of Results: mean/median (SD) Relevant Quality year diabetes sample measurement conclusions assessment sample rating†

No. + mean No. + mean Physical Sedentary age (SD) age (SD) activity behaviour

Objective method of PA/SB measurement (continued from previous page)

Loprinzi USA D (n=407). ND (n=1346). ActiGraph 7164 Light intensity Percent of day Participants Good (2014)13 Mean age Mean age accelerometer. PA: spent with T2D, 73.4 74.3 T2D 31.5% sedentary: compared to Step count PA ND 34.6% D 67.6% ND, engaged classification: ND 64.2% in significantly Sedentary: MVPA: more 0–99cpm D 0.8% sedentary Light: ND 1% behaviour and 100–2019cpm less light PA Moderate: Meeting PA and MVPA. 2020–5998cpm guidelines: Vigorous: D 9.1% No ≥5999cpm ND 13.1% differentiation between T1D and T2D

Steeves USA T2D (n=270). ND (n=302). ActiGraph Mean – Adults with T2D Fair (2015)14 Mean age Mean age AM-7164 accelerometer accumulated (SD) (SD) accelerometer TAC. less PA 70.6 (0.5) 69.8 (0.6) T2D : compared to weekdays ND 155 520; weekend days 140 302. ND: weekdays 202 011; weekend days 187 861

Van der Berg Netherlands T2D (n=714) ND (n=1395) ActivPAL3 Waking day Waking day Adults with Good (2016)15 Mean age Mean age accelerometer spent stepping: spent sedentary T2D spent 26 (SD) (SD) T2D 10.7% T2D 64.5% minutes more 62.7 (7.7) 58.1 (8.1) ND 13.6% ND 57.6% per day sedentary than those with ND

Hamer United T2D (n=112). – Actigraph GT3X Light mean Min/day spent – Good (2014)16 Kingdom Mean (SD) accelerometer. (SD): 186.9 sedentary age 63.9 (7) Step count PA (70.7) min/day. 659.6 (111.6) classification: MVPA mean Sedentary: (<1.5 (SD) 32.1(23.2) Proportion of MET) 0–199cpm. min/day day spent in Light: (1.5–3 sedentary MET) 200– Proportion of 75.1% 1998cpm. day spent in: MVPA: (>3 MET) Light activity ≥1999cpm 21.3%. MVPA 3.6%

Appendix 1. Study characteristics and quality assessment of included research. (Continued from previous page and continued on next 5 pages)

First Country Type 2 Comparison Method of Results: mean/median (SD) Relevant Quality author, diabetes sample measurement conclusions assessment year sample rating†

No. + mean No. + mean Physical Sedentary age (SD) age (SD) activity behaviour

Subjective methods of PA/SB measurement

Cassidy United T2D ND IPAQ-SF Not meeting PA >3 hrs/day Adults with Fair (2016)17 Kingdom (n=4074). (n=103 993). recommendations: watching TV: T2D report less 37–63 years 37–63 years ND 16.1% ND 20.3% PA and more of age of age T2D 24.7% T2D 38.1% SB than those without T2D

Sibai USA T2D (n=187). ND (n=2008). IPAQ-LF (IPAQ Low: – Adults with Poor (2013)18 ≥25 years of ≥25 years of scoring protocol T2D 55.6% T2D report less age age used*) ND 45.9% PA than those Moderate: without T2D T2D 34.8% ND 30.7% High: T2D 9.6% ND 23.4%

Karjalainen Finland T2D ND Self-administered Vigorous PA: – Women, but Fair (2008)19 (n=195). (n=1750). questionnaire T2D 56% not men, with Mean (SD) age Mean (SD) age (researcher ND 63% T2D self-report 64.3 (7) 58.3 (8.2) developed PA compared questions): Lifestyle PA ≥30 to those min/day: without T2D Vigorous PA ≥2–3 T2D 49% times a week for ND 57% 20–30 min

Lifestyle PA ≥30 min/day

Morrato USA Diabetes ND The Medical Physically active: – Adults with Fair (2007)7 (n=1825). (n=21 401). Expenditure Panel diabetes Age 18+ Age 18+ Survey (MEPS) D 39% of significantly years (no years (no participants less PA than upper age upper age ‘Physically active’ adults without limit) limit) defined as: ≥30 ND 58% of diabetes. mins in MVPA ≥3 participants times a week Data adjusted for sex, age, race/ethnicity, education and income levels, region, BMI, cardiovascular risk factors, depression, physical limitation status

No differentiation between T1D and T2D

Appendix 1. Study characteristics and quality assessment of included research. (Continued from previous page and continued on next 4 pages)

First author, Country Type 2 Comparison Method of Results: mean/median (SD) Relevant Quality year diabetes sample measurement conclusions assessment sample rating†

No. + mean No. + mean Physical Sedentary age (SD) age (SD) activity behaviour

Subjective methods of PA/SB measurement (continued from previous page)

Zhao USA D: ND: BRFSS Survey Participants – People with Poor (2008)20 2001 2001 meeting PA diabetes were less (n=5869); (n=116 984); PA based on ADA guidelines likely to meet either 2003 2003 2007 guidelines mean (SD). national or ADA (n=8888); (n=144 533); recommendations 2005 2005 Diabetes: for PA than people (n=13 608) (n=191 369) without diabetes. 2001: Age 18+ years Age 18+ years 38.5 (1.5)% The percentage of (no upper age (no upper age people with limit) limit) 2003: diabetes who 39 (1.5)% participated in PA in the past 10 years or 2005: met PA 41.7 (1.6)% recommendations in the past 5 years did No diabetes: not vary, whereas significantly 2001: increasing trends 46.6 (0.2)% were observed in people without 2003: diabetes. 47.7 (0.2)% Multivariate 2005: adjustment made for 49.8 (0.2)% age, sex, BMI, race/ ethnicity, education, income, smoking and employment/ work status.

No differentiation between T1D and T2D

Zhao USA D ND BRFSS Survey Meeting PA – Adults with diabetes Poor (2011)21 (n=18 370) (n=80 802) guidelines: significantly less PA 19.8% of PA based on ADA than adults without total sample 2007 guidelines Diabetes 25% diabetes. ND 34% Mean age of Data stratiﬁed total sample according to age, 74.3 years sex, and race or ethnicity.

No differentiation between T1D and T2D

Appendix 1. Study characteristics and quality assessment of included research. (Continued from previous page and continued on next 3 pages)

First author, Country Type 2 Comparison Method of Results: mean/median (SD) Relevant Quality year diabetes sample measurement conclusions assessment sample rating†

No. + mean No. + mean Physical Sedentary age (SD) age (SD) activity behaviour

Subjective methods of PA/SB measurement (continued from previous page)

Babić Croatia T2D (n=76). ND (n=60) Baecke’s Sport Index:** – Lower PA in all Poor (2004)22 20–65 years 20–65 years questionnaire and ND 2.05 groups of men of age. of age. LRCPA T2D 1.86 and obese Mean (SD) Mean (SD) questionnaire women with age 52 (10) age 52 (10) T2D compared to general population

Duarte Brazil T2D (n=118). T1D (n=107). IPAQ-LF T2D: – Larger Fair (2012)23 Low 30.7% percentage of Mean (SD) Mean (SD) (IPAQ scoring Moderate T2D classified age 62 (11) age 37 (11) protocol used*) 60.6% as poorly active High 8.7% and a lower percentage T1D: classified as Low 10.3% highly active Moderate compared with 64.7% T1D High 25%

29.7% of adults with T2D reported exercising regularly compared with 46.7% of adults with T1D

Plotnikoff Canada T2D T1D (n=697). GLTEQ MVPA in – Adults with T2D Poor (2006)24 (n=1614). MET-min/wk, report lower <600 MET-min/ mean (SD): levels of PA Mean age Mean age wk = T1D 677.6 compared to (SD) (SD) ‘inadequately (1136.3) adults with T1D 62.9 (12.1) 51.1 (17.1) active’ T2D 523.3 (1197.3)

Inadequately active: T1D 63.7% T2D 71.9%

Thomas United T2D (n=329). T1D (n=77). Questionnaire Physically – No direct Poor (2004)25 Kingdom (researcher active comparison Mean age Mean age developed participants: between T2D (SD) (SD) questions) T2D 32% and T1D 60.5 (11.9) 33.9 (8.4) T1D 39% reported

Çolak Turkey T2D (n=129). – IPAQ-SF Low 39.5% Mean sitting – Fair (2016)26 20–79 years Moderate duration: of age. (IPAQ scoring 51.9% 302min/wk Mean (SD) age protocol used) High 8.5% 56.4 (11.1)

Appendix 1. Study characteristics and quality assessment of included research. (Continued from previous page and continued on next 2 pages)

First Country Type 2 Comparison Method of Results: mean/median (SD) Relevant Quality author, year diabetes sample measurement conclusions assessment sample rating†

No. + mean No. + mean Physical Sedentary age (SD) age (SD) activity behaviour

Subjective methods of PA/SB measurement (continued from previous page)

Cooper USA T2D (n=253). – IPAQ-SF + Low 62.9% Mean (SD) – Fair (2015)27 Summary of Moderate 15% sitting hours Mean (SD) Diabetes Self-Care High 22.1% per day age 57.93 Activities Measure 5.82 (4.53) (11.52) (IPAQ scoring protocol used)

Nolan Australia T2D (n=408). – IPAQ-SF 57% PA – – Fair (2016)28 Mean age (SD) recommendations 62.7 (8.6)

Oguntibeju Nigeria T2D – IPAQ-SF Low 62% – – Poor (2012)29 (n=100). Moderate 34% Between (IPAQ scoring High 4% 40–85 years protocol used*) of age

Ojoawo Nigeria T2D – IPAQ-SF Low 26% – – Poor (2012)30 (n=100). Moderate 41% Between (IPAQ scoring High 33% 40–85 years protocol used*) of age

Cloix France T2D (n=724). – IPAQ-LF Low 15.1% – – Fair (2015)31 Mean (SD) Moderate 51.3% age 62 (10) (IPAQ scoring High 33.6% years protocol used*)

Kabanda South Participants – IPAQ-LF 39.7% of – No Poor (2011)32 Africa with DM participants differentiation (n=156). Participants who were inactive between T1D Mean age accumulated and T2D 48.7 years <599 MET-min/ wk were classified as inactive

Oyewole Nigeria T2D (n=119). – IPAQ-LF 31.1% of Mean sitting – Fair (2014)33 Mean age participants time: (SD) Participants who were inactive 288min/day 61.8 (11.8) accumulated <599 MET-min/ wk were classified as inactive

Appendix 1. Study characteristics and quality assessment of included research. (Continued from previous page and continued on next page)

First author, Country Type 2 Comparison Method of Results: mean/median (SD) Relevant Quality year diabetes sample measurement conclusions assessment sample rating†

No. + mean No. + mean Physical Sedentary age (SD) age (SD) activity behaviour

Subjective methods of PA/SB measurement (continued from previous page)

Hays USA T2D (n=260). – Interviewer Weekly PA – – Poor (1999)34 Mean age 67 administered reported: years survey (researcher 1–60 minutes developed 22% questions) >60 minutes 23%

55% of participants reported 0 minutes of PA per week

Parajuli Nepal T2D (n=385). – 3-day recall Poor adherence – – Poor (2014)35 Mean age questionnaire to PA 36%; (SD) (researcher Good 54.4 (11.5) developed adherence to questions using PA 21%. compendium of PA) 42.1% of participants were inactive

Appendix 1. Study characteristics and quality assessment of included research. (Continued from previous pages)