You are on page 1of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Title: Engaging self-correcting feedback control to increase physical activity and reduce bodyweight and disease risk in overweight sedentary adults.

Running Title: Self-correcting feedback control for weight loss

Authors: Kraushaar Lutz Erwin*, MSc., Department of Public Health Medicine, School of Public Health, University of Bielefeld, POB 100131, 33501 Bielefeld, Germany, lutz.kraushaar@uni-bielefeld.de Krmer Alexander, M.D., Ph.D., Professor and Head, Department of Public Health Medicine, School of Public Health, University of Bielefeld, POB 100131, 33501 Bielefeld, Germany * Corresponding author

Date:

03/2010

This study was supported by Siemens Betriebskrankenkasse (SBK), Siemensallee 84, 76187 Karlsruhe, Germany

20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

Abstract Objective: To investigate whether engaging web-enabled cognitive feedback control over the introduction of leisure time physical activity (LTPA) will yield adoption of health enhancing LTPA volumes among sedentary, overweight adults, and promote clinically relevant improvements of anthropometric, metabolic and fitness-related vital signs. Design: Longitudinal LTPA intervention study, commencing with a minimum weekly requirement of 3x20 minutes of high-intensity interval training (HIT), and requirement for web-based self-monitoring and reporting of LTPA volume and body weight. Subjects: 83 overweight, sedentary, otherwise healthy adults (age 26-68y, BMI 25.1-. 41.7 kg/m2, 24% female). Measurements: Anthropometric parameters, body fat (phase sensitive multi-frequency BIA), total-to-HDL cholesterol ratio, VO2peak (cardiopulmonary exercise testing, CPET), self-reported time spent for LTPA, frequency and latency of use of the web-enabled tool. Results: At 24-week follow-up, substantial voluntary increase of time spent for LTPA (mean and median of 135 and 170 minutes per week respectively) in the group of 72% of participants who successfully engaged cognitive feedback control (CFG), vs. no increase in the remaining participants who served as the control group (CG). CFG witnessed significantly improved peak oxygen consumption >1 metabolic equivalent (MET) vs. no improvement in CG. CFG also reduced BMI, body weight, body fat and TCH/HDL by 1.6 kg/m , 4.8 kg, 3.6 kg and 0.25 respectively in CFG vs. 0.4 kg/m , 1.4 kg, 1.1 kg and an increase in TCH/HDL ratio (0.16) in CG. Conclusion: Engaging selfcorrecting feedback via internet-based self-monitoring and feedback control may be a promising strategy for instituting sustainable health enhancing behavior change in overweight adults, offering the possibility of open-ended intervention delivery at low costs.
2 2

Key words: Obesity, Physical Activity, Individual Behavior, Homeostasis, Feedback

44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74

Background Excess body weight causally contributes to the development of cardiometabolic disease [1]. That leaves the question how to institute lasting weight loss and weight maintenance in individuals whose physical activity and dietary habits have promoted the onset of overweight and obesity. Lifestyle interventions, which target participants deficits in LTPA and their surfeits in caloric intake, may perform well in initiating weight loss and improvements of disease risk. However, an almost complete reversal to baseline status within a 3-5 years post-intervention period has been observed [2, 3], as well as a 95% failure rate of dieters attempts at losing weight and maintaining weight loss in the long term [4]. These data suggest that a sustainable public health strategy for health behavior change has yet to be found. To this end we developed a biobehavioral model with which to explain the observation of runaway weight gain in our society, and from which to formulate a testable hypothesis for sustainable remedial intervention. Since intervention efficiency for public health is our objective, we desired the intervention to allow for optimum reach into the at-risk population under the given economic, regulatory and resource constraints which define the local German health care system. The biobehavioral origin of the obesity epidemic The ecological observations guiding the development of our hypothesis were (a) the absence of excess body weight in human societies of hunter/gatherers living in their natural habitat [5, 6], (b) the obligatory physical activity cost for food acquisition in this habitat [6, 7], and (c) mans progressive weight gain secondary to the abolition of obligatory energy expenditure [7] and the introduction of processed foods of high energy density in modern society. We posit that the latter derails the anabolic and catabolic constituents of a negative feedback loop, which autonomously controls energy homeostasis. Figure 1 presents our proposed feedback loop model, of which three essential aspects warrant further elaboration. Firstly, the model positions catabolic foraging and anabolic feeding as the inextricable appetitive and consummatory components of ingestive behavior (AIB & CIB) [8, 9]. Their underlying neurohormonal pathways autonomously establish energy homeostasis through negative feedback control [10]. The hormones neuropeptide Y (NPY) and leptin have emerged as the chief stimulator and moderator respectively of an organisms drive to acquire food [11-14], which has been found to operate with remarkable similarity in all vertebrates and even in some fish [15, 16]. Secondly, there is dopamine as the neural substance that actuates the wanting for hedonic experiences, which an organism has learned to associate with certain stimulants [17]. Sweet and fatty tastes are such stimulants which fuel mans dopaminergic drive [18] and consequently his cravings for 3

75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105

sweet and fatty foods [19]. The drive to preferentially select such comestibles and their caloric content may have carried a distinct survival value for an organism subsisting in a habitat which is characterized by volatile food supplies and constant demands for physical activity. In modern human society however, the addictive power of typically sugar- and fat-enriched processed foods conspires with the abolished need for physical activity to derail energy homeostasis and its underlying neurohormonal system of negative feedback control. A functioning negative feedback control implies that once an individuals drive to forage is activated, and food reward is presented free of the energy cost of preceding PA, the eventual strengthening of the leptin signal, secondary to an accumulation of energy reserves, would moderate any subsequent motivation to forage when the latter is initiated by energy flux activated NPY signaling. However, at least two conditions have been observed which potentially derail negative feedback control. First, there is leptin resistance, a condition frequently encountered in overweight individuals [20]. Second, there is conditioned potentiation of feeding, a variant of classical conditioning, in which a previously unrelated stimulus, when paired with food presentation, arouses the dopamine driven wanting of food when the then conditioned cue and access to food are paired subsequently [21]. Once stimuli, such as time of day, physical location or presence of others have been conditioned into cues for food intake, a dopaminergic drive is engaged to eat in excess of physiological need. In our model, the controlled parameter of negative feedback is energy adequacy under the given environmental constraints, rather than bodyweight, as suggested by set-point theory [22]. The latter is challengeable on observational and evolutionary grounds. Its prediction of eventual weight stability fails to reconcile with the observation of longitudinally increasing bodyweights of societies and of individuals. The ability to cap body weight would have increased inclusive fitness only if environmental conditions had facilitated pathological weight increase to the point of affecting an individuals chances to reproduce and survive. This is hardly reconcilable with our current understanding of the scarcity and volatility of food supplies that has characterized the hominid environment throughout evolution. In defense of the theory, the failure of set-point control has been blamed on a purely cognitive/executive decision-override of homeostatic body weight control [23], which amounts to blaming the overweight individual for his predicament. Contrary to this view, our model exonerates the overweight individual, as it suggests his behavior to be driven by evolutionary conditioned neurohormonal mechanisms, which are autonomous in nature [10] and maladapted to the challenges of the modern environment. 4

106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136

In this context, voluntary LTPA offers itself as a cognitively controllable means to correct the energy imbalance and metabolic consequences, which result from the absence of the obligatory physical activity cost of food in the industrialized human society. In support of this view, LTPA has been presented as the critical component of sustained weight loss in the long-term follow-up of successful weight reducers [24, 25]. However, interventions aimed at increasing LTPA suffer from high attrition rates of typically 50% within the first 6 months [26-28]. As an explanation we offer (a) failure to consider participants time constraints, and (b) failure to engage feedback control. With respect to time constraints, lack of time is sedentary individuals most frequently cited obstacle to the cultivation of a regular exercise habit [29-31]. This real or perceived lack of time emerges relative to current recommendations, such as those of the American College of Sports Medicine [32] or the Institute of Medicine [33] who advocate 150-250 minutes per week and 60 minutes per day respectively of health enhancing physical activity (HEPA). Failure to overcome the discrepancy between time required and time perceived available for HEPA may inadvertently either freeze individuals into their sedentary habits, or promote recidivism from attempts at adopting HEPA. With respect to cognitive feedback control, its sine qua non is self-monitoring, which, when practiced, has been found to significantly improve adherence to behavior change [34]. The hypothesis Taken together, the observations discussed above led to our hypothesis that engaging self-monitoringbased feedback control over an initially minimal but acceptable LTPA volume will promote the voluntary adoption of progressively increasing PA volumes in previously sedentary overweight adults, who self-selected for participation in a weightloss intervention. We further hypothesized that selfmonitored and self-reported PA volume will correlate with objectively measurable vital signs of body weight and physical fitness. Overcoming the economic obstacles to effective lifestyle change The high costs of evidence-based lifestyle interventions, health care providers time limitations and a lack of reimbursement for health care providers preventive services constitute substantial barriers to the provision of preventive lifestyle change interventions [35]. In Germany, legislation mandates health insurers to subsidize, but not fully reimburse, members voluntary participation in selected exercise and diet programs, up to an annual ceiling. However, utilization of these funds typically is neither targeted to the beneficiarys health profile nor is it informed by medical advice or guidance. This strategy renders the preventive efforts efficiency probably sub-optimal. An internet-based intervention, 5

137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167

which is (a) targeted to an individuals health profile, which (b) automates, standardizes and maintains the process of engaging feedback control over HEPA, and which (c) integrates into the statutory and economic landscape of the health care system could constitute an economically viable and evidencebased alternative. To this end we consulted with a medium-sized statutory health insurance provider with the aim of operationalizing the intervention to meet with the acceptance of the primary cost carrier of health care services in Germany. At a 12.- monthly deductible, to be contributed out-of-pocket by the participants, the insurer considered the proposed intervention to satisfy economic and statutory constraints, and subsequently agreed to its realization as a pilot project within the community setting of an industrial estate in the South-Western German city of Karlsruhe. The 12.- out-of-pocket benchmark had emerged from an evaluation of primary care patients willingness-to-pay for preventive services, conducted in a primary-care setting of the close-by community of Heidelberg [36]. Of the 967 survey respondents (99.4% response rate) recruited from among 5 primary care practices, 27%, 40% and 12% had indicated willingness-to-pay less than 15, 15 - 40, and >40 respectively for preventive services. Methods The intervention was designed as a non-randomized controlled trial in sedentary and overweight, apparently healthy adult men and women. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki. Approval was obtained from the ethics committee of the state medical board of Baden-Wrttemberg. All participants gave written informed consent prior to enrollment. Subjects Subjects were recruited from among 200 German holders of a compulsory health insurance policy who had taken up their insurers invitation to participate in a subsidized fitness and physical activity examination. All subjects were employees at a large industrial estate of a multinational German electronics manufacturer. Figure 2 presents an overview of the recruitment process. Inclusion criteria were a self-reported current volume of LTPA of 1 hour or less per week and a body mass index (BMI) in excess of 25 kg/m . Exclusion criteria were known diseases and physical disabilities preventive of participation in an exercise program. All participants were Caucasians of German extraction. Intervention The minimum requirement for all participants was an unsupervised exercise protocol of HIT of thrice weekly 20 minutes (either running or cycling) in line with U.S. government recommendations, 6
2

168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198

published under the Healthy People 2010 initiative, and calling for moderate-intensity exercise of at least 30 minutes on at least 5 days per week, or alternatively, for 20 minutes high-intensity exercise at least thrice weekly [37]. Thrice weekly 15-minute HIT bouts have shown to yield significant improvements of parameters of metabolism and exercise capacity [38]. Each 20-minute HIT session was to consist of 4 repeated 60-s sprints at a heart rate commensurate with 85% to 95% of participants individual VO2peak with a 4-min recovery phase between sprints. During recovery, subjects were to continue their mode of exercise at an intensity level commensurate with their anaerobic threshold. Subjects were instructed not to perform HIT on consecutive days, but were encouraged to additionally engage in moderate-intensity endurance training at 95-115% of their individual anaerobic threshold. All participants who opted for the use of heart rate monitors during exercise were given target heart rates for the HIT exercises and recommendations for the optimal heart rates during continuous aerobic exercise. All heart rate recommendations were based on the individuals cardiopulmonary exercise test results. Participants who decided against the use of heart rate monitors were familiarized with the use of the 10-point OMNI rating scale of perceived exertion [39] and instructed to perform the highintensity intervals at an approximate rating of 8 and the recovery phase at a rating of 5-6. The OMNI scale has been validated for use in equivalent populations [40, 41]. Self-Monitoring To facilitate self-monitoring, and the supervision thereof by the investigator, an electronic lifestyle file (ELF) was created into which participants were to report their actual time spent on exercise and their bodyweight. The ELF facilitates a 6-weeks cumulative graphical display of actual vs. target values. Target performance for the initial 6-weeks period was based on the weekly 60-minutes HIT protocol. Upon completion of each 6-weeks period the target for the following 6 weeks was set to increase by 10% over the actual volume reported for the preceding 6-week period. Participants were encouraged to log their actual PA performance and their bodyweight on a daily basis, either by direct access to their secured web-page or through an applet installed on their mobile phone facilitating SMS-based reporting of PA and bodyweight. Measurements Body weight and standing height were measured in light sports clothing and without shoes to the nearest 0.1 kg and 1 cm, using a wall-mounted anthropometer and a calibrated electronic scale, respectively. BMI was calculated as the ratio between weight and height squared (kg/m ). 7
2

199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229

Body composition was measured using an impedance analyzer device and software (BIA 2000-S, Data Input, Frankfurt, Germany) for tetrapolar bioelectrical impedance analysis (BIA) measurement of resistance (R) and reactance (Xc) at frequencies of 5, 50 and 100 kHz. Measurements were made at the right side of the subject between the wrist and ankle while in a supine position and after having rested for 5 minutes. The equipment, analytic algorithms and the measurement protocol have been validated previously in comparable populations [42, 43]. Exercise testing was performed as cardiopulmonary exercise test on a cycle ergometer (Customed, Germany) using a ramp protocol to exhaustion with the ramp increment chosen, based on age, weight, height and training history, as to reach exhaustion within 8 to 12 minutes [44]. For the first 3 minutes the workload was fixed at 5 W. The resistance on the cycle ergometer was controlled by the ergospirometric software (Cortex, Leipzig, Germany) to be independent of pedal cadence. Spiroergometry was carried out using a breath-by-breath-system (Cortex MetaLyzer 3B, Leipzig, Germany), which has been validated previously [45]. Expired air was collected continuously using a facemask. The system was calibrated prior to each test in accordance with manufacturers guidelines using a 3-L syringe for volume calibration and ambient air measure for gas calibration. During all tests, heart rate was recorded with a wireless chest strap telemetry system (Polar, Kempele, Finland). Simultaneous gas exchange measurements consisted of minute ventilation (VE), oxygen uptake (VO2; electrochemical cell), and carbon dioxide output (VCO2; infrared analyzer). For calculations, data were averaged over every 20 seconds. Peak oxygen uptake (VO2peak) was defined as the highest value for oxygen uptake averaged over 20 seconds. Venous blood was sampled in EDTA tubes in the morning between 07:30 and 08:45 after an overnight fast. Total and HDL cholesterol were determined by standard laboratory methods using certified assays in a local clinical laboratory. All analyses performed at baseline were repeated at follow-up. Adherence Definition Adherence was defined as meeting the minimum criteria of having recorded a minimum weekly duration of endurance exercise (volume aspect of adherence) of 60 minutes (3 x 20 minutes of HIT) or more for at least 12 consecutive weeks (duration aspect of adherence), with the last self-reported login not earlier than 1 week (latency aspect of adherence) prior to the date of final assessment. With 8

230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260

physical fitness being the primary and measurable vital sign, the 12-weeks duration is in keeping with published evidence, which suggests that measurable effects accrue to VO 2peak after such durations [46], with decay of the effect being observable within 14 days of discontinuation of the exercise regimen [47]. Participants who did not meet the adherence criteria were considered the control group. Statistical Analyses Prior to the study, we performed power and sample-size calculations, with both calculations based on a hypothesized ratio between adherent and non-adherent participants r=3. To achieve a power of 90%, we needed 50 participants to detect a between-group difference for VO2peak of 1 MET (3.5 ml/kg/min) and for BMI of 1 kg/m . Analyses for differences between groups at baseline were performed using t-tests. Changes from baseline to follow-up were tested using paired t-test for withingroup changes, and unpaired t-tests for differences of changes between groups. Statistical significance was accepted at p<0.05. Analyses were carried out with Intercooled STATA 11 for Macintosh (Stata Corp. Texas, U.S.A.) and all data are expressed as means 1SD unless noted otherwise. Results Of the 83 enrolled participants 82 were available for follow-up analyses. The one unavailable participant had also failed to meet the adherence criteria. Adherence rate was 72.3% (95% CI: 62.6% 81.9%) with 60 individuals having met the adherence criterion at follow-up and therefore being considered the group which successfully engaged consciously controlled feedback (CFG). The remaining 23 participants are considered the control group (CG). of the 60 adherent participants, two had refused participation in venous blood sampling at follow-up. Table 1 summarizes the relevant parameters at baseline and their changes to follow-up for both groups. There were no significant differences at baseline between CFG and CG for any of the variables. At 24-weeks follow-up, CFG witnessed a significant improvement of VO2peak of 3.7 ml/kg/min vs. a non-significant decrease of 0.6 ml/kg/min among CG participants. VO2peak is conventionally expressed in milliliters of oxygen consumption per minute and per kilogram of bodyweight. Hence, any post-weight loss increase in VO2peak could be merely an arithmetic artifact without any concomitant real increase of CRF. Since it is a bodys lean mass (LBM), not its fat content, which delivers the muscular work and power output, the correlation of VO 2 consumption with LBM provides a more adequate value to assess the true change in cardiopulmonary fitness. For this purpose we examined VO2peak in relation to LBM. While CG witnessed a non-significant decrease of 9
2

261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291

VO2peak of 2.4 ml/kgLBM/min, CFG had significantly increased VO 2peak by 6.2 ml/kgLBM/min, with p-value for within- and between-groups of <0.001 respectively. CFG significantly reduced BMI, body weight and body fat, whereas the decrease of BMI in CG was borderline significant at p=0.5 but significant for body weight and body fat. However, the decreases in anthropometric variables among CFG participants exceeded those of CG participants by a factor of >3, with a significance for this between-group difference at p<0.001. Figure 3 shows the relative changes witnessed in both groups for VO2peak, BMI, body weight and body fat. CFG had a significantly decreased TCH/HDL ratio by 0.25 at follow-up vs. a non-significant increase 0f 0.16 in CG. The combined between-group difference of 0.41 was significant two-tailed, with t(78)=2.9, p<0.01. In CFG the median and mean weekly reported time spent for exercise was 135 minutes and 170 minutes respectively, being more than twice the minimum requirement for the definition of adherence, and twice the values of 58 minutes and 81 minutes achieved in CG as the median and mean respectively. Discussion In this study we show that overweight individuals, enrolled into a PA-based weight loss intervention, voluntarily increase their PA volume over an initial minimal-dose requirement to yield significant and clinically relevant improvements of body weight status, lipid parameters and measures of physical fitness. The 1-MET increase of VO2peak in CFG is of clinical relevance, as cardiorespiratory fitness (CRF) has been found to linearly correlate, at 1-MET increments, with a decreased risk of cardiovascular disease (CVD) events in a population of similarly aged Finnish men [48]. Similarly, a risk-reduction of 8-12% per each 1-Met increase of CRF has been reported by others [49]. The reduction of body weight by 5% and less than 1.5% in CFG and CG respectively bespeaks a clinically relevant effect, given that a minimum weight loss of 5% has been suggested to be required for clinically relevant hormonal improvements [50], with others arbitrarily setting the bar somewhat lower at 3% for the definition of successful long-term maintenance of weight loss [51]. The fact that approximately of the achieved weight loss originates from a reduction of body fat, bespeaks a clinically desirable effect, as it is the hormonal products of fat tissue which exert the deleterious effects of excess weight on physical health [52, 53]. In this study we also show that it is possible to engage a cognitively controlled feedback loop, by using 10

292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322

a web-based self-monitoring tool, to institute physical activity in previously sedentary and overweight adults. The results of this study support the hypothesis that by engaging self-correcting feedback, with weight-loss being its objective, users voluntarily and significantly increase their physical activity volume over an initially prescribed minimum. To start with a minimum prescription may be an important strategy to overcome sedentary individuals perceived gap between time required and time available for HEPA. Engaging a cognitively controlled self-correcting feedback loop may therefore help previously sedentary and overweight adults to voluntarily, gradually and substantially increase their physical activity volume over an initially prescribed minimum, resulting in significant reductions of body weight and associated disease risk. A call has been made very recently for the EU to develop national physical activity recommendations along the new guidelines formulated by the U.S. American Heart Association (AHA) and the American College of Sports Medicine (ACSM) [54]. These guidelines specifically acknowledge the evidence-based need for all healthy adults aged 18-65 to perform either moderate-intensity aerobic exercise for a minimum of 30 min five times weekly, or 20 min of vigorous exercise 3 times weekly or any combination thereof [55]. This interventions starting point was a CPET-based heart rate-controlled exercise recommendation for an initial 3x20 minutes HIT routine. Individualizing an exercise prescription to a participants personal cardiopulmonary dynamics maximizes the training effect. The hoped-for consequence is that a perceivably improved performance powerfully promotes continued adherence. Letting participants self-determine their PA volume, given the initial weekly requirement of 3x20 minutes of HIT, yielded a substantially larger actually performed volume of exercise. It is therefore tempting to suggest, that it may not so much be participants fidelity to a pre-conceived one-size-fits-all exercise curriculum, which lifestyle change program providers should be concerned with. Rather should we focus on getting people to commit to and commence with a physical activity habit, which, if prescribed individually to yield some quick tangible effect, will develop its own momentum. It is for future research to contrast these two strategies under the a priori hypothesis of a significant difference in outcome. In this study, the term adherence does not relate to the volume or duration of PA, but to the presence of a cognitively controlled feedback loop at the end of a 6-months observation period. This constitutes a small, but substantially different way of defining adherence, which emerges from the theory and model underlying this intervention strategy. Its latency aspect differentiates it from adherence definitions, which are exclusively based on any combination of percentages of volume, of duration or 11

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353

of attendance. It provides an answer to the question, what proportion of study participants currently adheres to the PA protocol and has done so for durations and at PA volumes, which are expected to yield tangible health benefits. Hence, the latency aspect is essential for determining how successful an intervention has been at releasing its participants with a modified health habit. With 72% of the participants meeting this criterion, the intervention compares favorably with the 55% (CI 0.39-0.72) adherence rate achieved in a comparable worksite intervention of a 24-weeks PA program consisting of 3 weekly 20-minutes high intensity aerobic workouts in addition to strength training [56], and with the 60% adherence rates typically reported for PA interventions. This difference in adherence is significant, two-tailed, at p(z=2.28)<0.05. The internet-enabled monitoring of participants self-monitoring facilitates an up-to-date assessment of adherence status and can be deployed to alert clinicians to participants whose latency of login may indicate adherence decay or drop-out. Thereby motivational support can be provided to those who need it, without wasting resources on those who do not. The power of self-monitoring in promoting intervention adherence has been confirmed in observations from the U.S. national weight loss registry. Its registrants, who present a mean 30 kg weight loss maintained over >5 years, have found selfmonitoring, a high level of PA and a low-calorie-low-fat diet as the three most important determinants of weight loss and maintenance of reduced body weight [25, 57]. Strengths and Limitations One major strength of our study is its simulation of a real-life implementation, designed to be economically acceptable to the provider of statutory health insurance under which 90% of the resident population is covered. Another notable strength is the studys objective measurement of changes in body weight, body composition, TCH/HDL ratio and physical fitness resulting from an internetdelivered intervention which was designed to engage cognitive feedback control over physical activity behavior. The primary limitation of our study is that (a) these results represent initial improvements of weight and fitness status, and that (b) the study design, which necessitated participants out-of-pocket contributions with the objective of simulating a future real-life implementation, prevented us from randomizing participants into a control and intervention group. The resulting selection of non-adherent participants as the control group may have led to selection bias. However, there were no significant between-group differences of any of the parameters measured at baseline. Expressed as Cohens d, the differences in parameters at baseline remained at d0.1, with the exception of age, the difference 12

354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376

of which had a Cohens d=0.3. Conventionally, effect sizes are considered small, medium and large for d values of 0.2, 0.5 and 0.8 respectively [58]. Also, self-report of PA is inherently subject to bias. Since the self-reporting technique did not enable participants to differentiate between levels of exercise intensity, the calorific equivalent of the reported exercise volume cannot be determined. Hence, no dose-response relationship could be established between PA volume and outcome. Also, subjects were free to record either daily or as and when PA was performed. Enforcing daily login might have reduced the number of subjects who failed to engage a cognitively controlled feedback loop. In a follow-up trial we are encouraging participants to perform daily monitoring not only of body weight and PA but also of dietary intake. We hypothesize that daily feedback will significantly improve adherence and outcome. Based on this hypothesis, the ELF is being further developed to include dietary monitoring and to facilitate telemetric monitoring of PA volume and intensity as well as telemetric monitoring of body weight, blood pressure, blood glucose and ECG. We encourage fellow researchers to avail themselves to this tool, and welcome all enquiries related to academic research. Conclusion This study demonstrates that a web-enabled engagement of cognitive feedback control enables sedentary and overweight individuals to voluntarily increase LTPA to yield clinically relevant improvements of anthropometric, metabolic and fitness related vital signs. The internet-enabled implementation provides for a low-cost open-ended intervention delivery to large at-risk groups, possibly facilitating sustainable improvements of health behaviors. Follow-up research should elucidate the determinants of sustainability and efficiency within the statutory and economic constraints of the given health care system.

13

377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 Conflict of Interest All authors declared to have no conflict of interest. Acknowledgement We thank all participants who took part in this study. Funding The study was supported by Siemens Betriebskrankenkasse (SBK). Contributors L.E.K. contributed to the design of the study, collection and assembly of the data, analysis and interpretation of data and drafting the article. A.K. took part in the interpretation of data and drafting of the article. All authors approved the final manuscript. L.E.K. accepts full responsibility for conducting the study.

14

392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 12. 11. 10. 9. 8. 7. 6. 5. 4. 3. 2. References 1. Lavie, C.J., R.V. Milani, and H.O. Ventura, Obesity and Cardiovascular Disease: Risk Factor, Paradox, and Impact of Weight Loss. J Am Coll Cardiol, 2009. 53(21): p. 1925-1932. Miller, W.C., How effective are traditional dietary and exercise interventions for weight loss? Med Sci Sports Exerc, 1999. 31(8): p. 1129-34. Wing, R.R., et al., Behavioral science research in diabetes: lifestyle changes related to obesity, eating behavior, and physical activity. Diabetes Care, 2001. 24(1): p. 117-23. Crawford, D., R.W. Jeffery, and S.A. French, Can anyone successfully control their weight? Findings of a three year community-based study of men and women. Int J Obes Relat Metab Disord, 2000. 24(9): p. 1107-10. O'Dea, K., CARDIOVASCULAR DISEASE RISK FACTORS IN AUSTRALIAN ABORIGINES. Clinical and Experimental Pharmacology and Physiology, 1991. 18(2): p. 85-88. Eaton, S.B., S.B. Eaton, 3rd, and M.J. Konner, Paleolithic nutrition revisited: a twelve-year retrospective on its nature and implications. Eur J Clin Nutr, 1997. 51(4): p. 207-16. Booth, F.W., et al., Waging war on modern chronic diseases: primary prevention through exercise biology. J Appl Physiol, 2000. 88(2): p. 774-787. Day, D.E., E. Keen-Rhinehart, and T.J. Bartness, Role of NPY and its receptor subtypes in foraging, food hoarding, and food intake by Siberian hamsters. Am J Physiol Regul Integr Comp Physiol, 2005. 289(1): p. R29-36. Baird, J.P., N.E. Gray, and S.G. Fischer, Effects of neuropeptide Y on feeding microstructure: dissociation of appetitive and consummatory actions. Behav Neurosci, 2006. 120(4): p. 93751. Qian, G. and L.H. Tamas, Neuronal control of energy homeostasis. FEBS letters, 2008. 582(1): p. 132-141. Grill, H.J. and J.M. Kaplan, The neuroanatomical axis for control of energy balance. Front Neuroendocrinol, 2002. 23(1): p. 2-40. Grill, H.J. and J.M. Kaplan, Interoceptive and integrative contributions of forebrain and brainstem to energy balance control. Int J Obes Relat Metab Disord, 2001. 25 Suppl 5: p. S73-7.

15

422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452

13.

Kalra, S.P., et al., Interacting Appetite-Regulating Pathways in the Hypothalamic Regulation of Body Weight. Endocr Rev, 1999. 20(1): p. 68-100.

14.

Ammar, A.A., et al., NPY-leptin: opposing effects on appetitive and consummatory ingestive behavior and sexual behavior. Am J Physiol Regul Integr Comp Physiol, 2000. 278(6): p. R1627-33.

15.

Larhammar, D. and E. Salaneck, Molecular evolution of NPY receptor subtypes. Neuropeptides, 2004. 38(4): p. 141-51.

16.

Cerda-Reverter, J.M., et al., Molecular evolution of the neuropeptide Y (NPY) family of peptides: cloning of three NPY-related peptides from the sea bass (Dicentrarchus labrax). Regul Pept, 2000. 95(1-3): p. 25-34.

17.

Robinson, T.E. and K.C. Berridge, The neural basis of drug craving: an incentive-sensitization theory of addiction. Brain Res Brain Res Rev, 1993. 18(3): p. 247-91.

18.

Berridge, K.C., Food reward: brain substrates of wanting and liking. Neurosci Biobehav Rev, 1996. 20(1): p. 1-25.

19.

Levine, A.S., C.M. Kotz, and B.A. Gosnell, Sugars and Fats: The Neurobiology of Preference. J. Nutr., 2003. 133(3): p. 831S-834.

20.

Scarpace, P.J. and Y. Zhang, LEPTIN RESISTANCE: A PREDIPOSING FACTOR FOR DIETINDUCED OBESITY. Am J Physiol Regul Integr Comp Physiol, 2008: p. 90669.2008.

21.

Seymour, B., Carry on Eating: Neural Pathways Mediating Conditioned Potentiation of Feeding. J. Neurosci., 2006. 26(4): p. 1061-1062.

22.

Harris, R.B., Role of set-point theory in regulation of body weight. FASEB J, 1990. 4(15): p. 3310-8.

23.

Berthoud, H.R., Multiple neural systems controlling food intake and body weight. Neurosci Biobehav Rev, 2002. 26(4): p. 393-428.

24.

Tate, D.F., et al., Long-term weight losses associated with prescription of higher physical activity goals. Are higher levels of physical activity protective against weight regain? Am J Clin Nutr, 2007. 85(4): p. 954-959.

25.

Wing, R.R. and S. Phelan, Long-term weight loss maintenance. Am J Clin Nutr, 2005. 82(1): p. 222S-225.

26.

Robison, J.I. and M.A. Rogers, Adherence to exercise programmes. Recommendations. Sports Med, 1994. 17(1): p. 39-52. 16

453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483

27.

O'Neil, H.A. and S.N. Blair, Enhancing adherence in clinical exercise trials. Quest, 2001. 53: p. 310-317.

28.

Graffagnino, C.L., et al., Effect of a community-based weight management program on weight loss and cardiovascular disease risk factors. Obesity (Silver Spring), 2006. 14(2): p. 280-8.

29.

Courneya, K.S., et al., A longitudinal study of exercise barriers in colorectal cancer survivors participating in a randomized controlled trial. Ann Behav Med, 2005. 29(2): p. 147-53.

30.

Ziebland, S., et al., Lack of willpower or lack of wherewithal? "Internal" and "external" barriers to changing diet and exercise in a three year follow-up of participants in a health check. Soc Sci Med, 1998. 46(4-5): p. 461-5.

31.

El Ansari, W. and G. Lovell, Barriers to exercise in younger and older non-exercising adult women: a cross sectional study in London, United Kingdom. Int J Environ Res Public Health, 2009. 6(4): p. 1443-55.

32.

Donnelly, J.E., et al., American College of Sports Medicine Position Stand. Appropriate physical activity intervention strategies for weight loss and prevention of weight regain for adults. Med Sci Sports Exerc, 2009. 41(2): p. 459-71.

33.

Brooks, G.A., et al., Chronicle of the Institute of Medicine physical activity recommendation: how a physical activity recommendation came to be among dietary recommendations. Am J Clin Nutr, 2004. 79(5): p. 921S-930.

34.

Stevens, V., Learning habits and skills, in The ABC of Behavior Change, J. Kerr, R. Weitkunat, and M. Moretti, Editors. 2005, Elsevier: Edinburgh.

35.

Whitlock, E.P., et al., Evaluating primary care behavioral counseling interventions: an evidence-based approach. Am J Prev Med, 2002. 22(4): p. 267-84.

36.

Wiesemann, A., et al., Patient willingness to pay for preventive measures in primary care: a study of five GPs in a German community. Soz Praventivmed, 2004. 49(4): p. 254-60.

37.

US Department of Health and Human Services, Objectives 22-2 and 22-3. In: Healthy people 2010 (conference ed, in 2 vols), US Department of Health and Human Services: Washington, DC.

38.

Babraj J, et al., Extremely short duration high intensity training substantially improves insulin action in young sedentary males BMC Endocrine Disorders, 2009. 9(3).

39.

Robertson, R.J., Perceived Exertion for Practitioners. Rating Effort With the OMNI Picture System. 2004, Champaign, Ill.: Human Kinetics. 17

484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513

40.

Utter, A.C., et al., Validation of the Adult OMNI Scale of perceived exertion for walking/running exercise. Med Sci Sports Exerc, 2004. 36(10): p. 1776-80.

41.

Robertson, R.J., et al., Concurrent validation of the OMNI perceived exertion scale for resistance exercise. Med Sci Sports Exerc, 2003. 35(2): p. 333-41.

42.

Bosy-Westphal, A., et al., Phase Angle From Bioelectrical Impedance Analysis: Population Reference Values by Age, Sex, and Body Mass Index. JPEN J Parenter Enteral Nutr, 2006. 30(4): p. 309-316.

43.

Bosy-Westphal, A., et al., Patterns of bioelectrical impedance vector distribution by body mass index and age: implications for body-composition analysis. Am J Clin Nutr, 2005. 82(1): p. 6068.

44.

Rolf F. Kroidl, S. Schwarz, and B. Lehnigk, Technische und formale Grundlagen, in Kursbuch Spiroergometrie. 2007, Thieme: Stuttgart. p. 16.

45.

Meyer, T., et al., Reliability of gas exchange measurements from two different spiroergometry systems. Int J Sports Med, 2001. 22(8): p. 593-7.

46.

Miyaki, A., et al., Effect of habitual aerobic exercise on body weight and arterial function in overweight and obese men. Am J Cardiol, 2009. 104(6): p. 823-8.

47.

Solomon, T.P.J., et al., Effects of exercise training and diet on lipid kinetics during free fatty acid-induced insulin resistance in older obese humans with impaired glucose tolerance. Am J Physiol Endocrinol Metab, 2009. 297(2): p. E552-559.

48.

Laukkanen, J.A., et al., The predictive value of cardiorespiratory fitness for cardiovascular events in men with various risk profiles: a prospective population-based cohort study. Eur Heart J, 2004. 25(16): p. 1428-37.

49.

Franklin, B.A. and D.P. Swain, New insights on the threshold intensity for improving cardiorespiratory fitness. Prev Cardiol, 2003. 6(3): p. 118-21.

50.

Varady, K., et al., Degree of weight loss required to improve adipokine concentrations and decrease fat cell size in severely obese women. Metabolism, 2009.

51.

Anderson, J.W., et al., Long-term weight-loss maintenance: a meta-analysis of US studies. Am J Clin Nutr, 2001. 74(5): p. 579-84.

52.

Festa, A., et al., The relation of body fat mass and distribution to markers of chronic inflammation. Int J Obes Relat Metab Disord, 2001. 25(10): p. 1407-15.

18

514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529

53.

Snijder, M.B., et al., What aspects of body fat are particularly hazardous and how do we measure them? Int. J. Epidemiol., 2006. 35(1): p. 83-92.

54.

Oja, P., et al., Physical activity recommendations for health: what should Europe do? BMC Public Health, 2010. 10(1): p. 10.

55.

Haskell, W.L., et al., Physical Activity and Public Health. Updated Recommendation for Adults From the American College of Sports Medicine and the American Heart Association. Circulation, 2007: p. CIRCULATIONAHA.107.185649.

56.

Atlantis, E., et al., Worksite intervention effects on physical health: a randomized controlled trial. Health Promot Int, 2006. 21(3): p. 191-200.

57.

Wing, R.R. and J.O. Hill, SUCCESSFUL WEIGHT LOSS MAINTENANCE. Annual Review of Nutrition, 2001. 21(1): p. 323-341.

58.

Cohen, J., Statistical power analysis for the behavioral sciences. 2nd ed. 1988, New York :: Academic Press.

19

530

Table 1: Physiological and Anthropometric Characteristics of the Participants


Baseline Mean SD Age1 (years) non-adherent (N=22) adherent (N=60) Gender1 (% female) non-adherent (N=22) adherent (N=60) VO2peak1 per kg body weight (ml/kg/min) non-adherent (N=22) adherent (N=60) VO2peak1 per kg lean body mass (ml/kg/min) non-adherent (N=22) adherent (N=60) TCH/HDL1 non-adherent (N=22) adherent (N=58) BMI1 (kg/m2) non-adherent (N=22) adherent (N=60) Body Weight1 (kg) non-adherent (N=22) adherent (N=60) Body Fat1 (kg) non-adherent (N=22) 27.5 9.0 -1.12 (-0.1 to -2.0) adherent (N=60) 27.5 8.4 -3.62; 4 (-2.6 to -4.7) Abbreviations: VO2peak = Peak Oxygen Consumption; TCH = total cholesterol; HDL = high-density lipoprotein cholesterol Data are mean SD unless otherwise specified. 1: p>0.05 for between-group difference at baseline; 2: significant difference from baseline to follow-up at p<0.05; 3: significantly different from non-adherent group at p<0.01; 4: significantly different from non-adherent group at p<0.001 92.6 91.6 13.1 13.2 -1.42 -4.82; 4 (-0.1 to -1.6) (-3.5 to -6.2) 29.7 29.8 3.7 3.5 -0.4 -1.62;4 (0 to -0.8) (-1.1 to -2.0) 4.10 4.42 1.12 1.06 +0.16 -0.252;3 (-0.13 to +0.46) (-0.38 to -0.11) 48 51 23 25 8 9 Mean Change 95% CI

32.2 32.3

8.01 8.10

-0.59 +3.732; 4

(-0.36 to +1.54) (+2.71 to +4.74)

84.6 86.3

14.2 16.6

-2.4 +6.22; 4

(-5.0 to +0.3) (+3.8 to +8.5)

531 532 533 534 535 536

20

537

538 539 540 541 542 Fig. 1. Negative Feedback Loop of Energy Homeostasis NPY = Neuropeptide Y

21

543 544 545 Fig. 2. Flowchart Recruitment

22

546 547 548 549 550 Fig. 3. Changes of vital parameters at 24-weeks follow-up. P-values centered in each bar refer to within-group changes from baseline to follow-up. P-values between the bars indicate significant between-group differences.

23