Introduction
Bariatric surgery is considered the most efficient intervention for morbid obesity, showing promising results on weight control and reducing comorbidities (
1,
2). However, 20–30% of the bariatric patients are not able to reach their targeted weight loss (
1,
3), while 60% of the bariatric patients are not able to maintain at least 30% of their weight loss on the long term (
4). This indicates a large diversity in effectiveness of bariatric surgery among patients, which persists after adjustment for different types of surgery (
3).
The diversity in effectiveness of bariatric surgery may relate to variances in adopting a physically active lifestyle after surgery. Physical activity has a critical role in weight loss and long-term weight maintenance (
5) and has a positive influence on body composition, cardiorespiratory fitness (CRF) and health-related quality of life (HRQoL) (
6‐
8). Although previous studies with short-term exercise interventions after surgery show improved outcomes (
9,
10), these well-controlled conditions do not reflect real-life setting including the effect of adopting an active lifestyle, which is necessary for long-term effectiveness (
11).
We previously showed that sports activity was positively associated to CRF whereas leisure time activity was positively associated to weight loss (
12). The present study builds further on these outcomes as we assessed the distinguished effects of either increasing or decreasing total PA on bariatric outcomes in a broader health spectrum. The aim of this study is to determine the relation between pre- to post-surgery changes in physical activity and weight loss, fat-free mass, CRF and HRQoL two years after surgery. We hypothesize that enhancing physical activity after bariatric surgery will contribute to the effectiveness of the surgical treatment on weight loss, body composition, CRF and HRQoL.
Methods
Study Population
Patient data with two years follow-up was prospectively collected in eight clinics of the Nederlandse Obesitas Kliniek (NOK), a Dutch Clinic that provides a comprehensive perioperative care program for patients with morbid obesity undergoing bariatric surgery. Patients were screened by an interdisciplinary team for eligibility for bariatric surgery based on the IFSO criteria (
13). Inclusion criteria consisted of (i) a primary laparoscopic Roux-en-Y gastric bypass (RYGB) or sleeve gastrectomy (SG) which was performed between March 2012 and May 2015, (ii) participation in the preoperative care program with at least two years follow-up and (iii) complete data on physical activity. Patients who underwent a revisional bariatric operation or a gastric band operation were excluded.
Perioperative Care Program
Before surgery, patients were enrolled in a 7-wk preoperative care program, provided by an interdisciplinary team consisting of a dietician, psychologist, physiotherapist and a physician, under the supervision of an internist and surgeon. The preoperative program consists of weekly group visits with an educational program on healthy lifestyle. After surgery, patients continue with a 15-month postoperative care program, with the aim to develop a healthy lifestyle without support of a bariatric care team. This program consists of group visits every three weeks, with each three consecutive 1-h sessions with a dietician, psychologist and physiotherapist. Furthermore, patients had regular individual medical check-ups with a physician and follow-up measurements (e.g. body composition, physical activity, CRF and HRQoL) up to 5 years post-surgery, to monitor the progress of the patient.
Measurements
Physical Activity
Self-reported physical activity (PA) was assessed by the Baecke questionnaire. The Baecke questionnaire was modified to improve validity in this population with morbid obesity. Questions about sweating were excluded since obesity is related to excessive sweating. Therefore, sweating may decrease after surgery due to weight loss while in the questionnaire a decrease in sweating is seen as a decline in PA (
14). Likewise, questions regarding the comparison of PA to peers were excluded, since obesity often occurs within specific layers of the community (
15,
16). Therefore, our population might show similar behavior as peers within their environment, which the questionnaire interprets as sufficient PA, while the overall PA level within this environment might be poor.
The modified Baecke questionnaire contained 12 questions with a 5-point Likert scale on how much time was spent on several activities. Subsequently, three index scores (range 1–5 points) were calculated for work, sports and leisure time. These index scores were combined into a total Baecke score with a range from 3 to 15 points, in which a higher score represents a higher PA level.
Weight Loss
Weight loss was expressed in both percentage of total weight loss (%TWL) and percentage of excess weight loss (%EWL). %TWL reflects weight loss with respect to total preoperative weight, while %EWL reflects the weight loss with respect to preoperative excess weight (= preoperative weight minus appropriate weight based on a BMI of 25 kg/m2). Therefore, %TWL and %EWL can be calculated by the following formulas:
$$ {\displaystyle \begin{array}{c}\% TWL=\frac{weight\ loss\ (kg)}{preoperative\ weight\ (kg)}\ast 100\\ {}\% EWL=\frac{weight\ loss\ (kg)}{\left( preoperative\ weight\ (kg)-\left(25\ast {\left( height(m)\right)}^2\right)\right)}\ast 100\end{array}} $$
Body Composition
Height and waist circumference were measured using a non-elastic measuring tape. Weight, fat mass, fat-free mass (FFM, which mainly consists of muscle mass), fat percentage and FFM percentage (%FFM) were measured by a bioelectrical impedance analysis (TANITA® brand, model BC-420MA) (
17).
Cardiorespiratory Fitness
Cardiorespiratory fitness (CRF) defined as VO
2max (i.e. maximal oxygen uptake) was estimated by the Åstrand test, a submaximal exercise test on a cycle ergometer (
18). In the first phase of the test, the resistance is increased every minute until a heart rate of ~120 bpm is reached. Then, the patient is instructed to continue cycling for six minutes at this resistance level, while the heart rate is continuously measured. After the test, the VO
2max is estimated by combining patient characteristics (age and sex) with the achieved power output and the heart rate during the last 30s of the test. The Åstrand test could not be performed if subjects weighed ≥182 kg or used beta blockers or tricyclic antidepressants. Estimated VO
2max was expressed in absolute VO
2max (L/min) and VO
2max relative to body weight (ml/kg/min) and FFM (ml/kg FFM/min). The absolute VO
2max represents the maximal rate of oxygen consumption that the body can transport and metabolize within e.g. a minute during exercise, reflecting the aerobic capacity. However, this maximal oxygen consumption is also dependent on body type, therefore relative VO
2max values takes body size into account to allow for comparison between individuals.
Health related quality of life (HRQoL) was determined by the RAND-36 questionnaire (
19). The questionnaire contains 36 questions divided over eight subscales: physical functioning, physical role functioning, emotional role functioning, vitality, mental health, social functioning, bodily pain and general health. All subscales are expressed in a score between 0 and 100 points, in which a higher score reflects a higher HRQoL.
Statistical Analysis
Statistical analyses were performed using SPSS 24 software (IBM SPSS Statistics for Windows, Version 24 IBM Corp., Armonk, NY, USA.). First, all continuous variables were visually inspected and tested for normality by the Shapiro-Wilk test. The population was divided into 3 groups based on their change in total Baecke score from pre-to post-surgery: a decreased PA group with >1 pt. decrease, a maintained PA group with a change between 1 and -1 pt., and an increased PA group with >1 pt. increase. Some practical examples of a > 1 point increase in total Baecke score are listed in supplemental Table 1. Pre-surgery characteristics were checked for differences between the groups by an ANOVA-test, Kruskal-Wallis test or Chi-Square test for continuous parametric-, continuous non-parametric- and categorical data, respectively. Then, pre- to post-surgery change in PA and health outcomes of the total population were determined by paired t-tests. Relation between change in PA and pre- to postoperative changes in outcomes (i.e. %EWL, %TWL, FFM (kg), %FFM, VO2max and HRQoL subscales) were assessed by an ANCOVA, adjusted for age, sex, type of surgery, preoperative BMI and preoperative physical activity. These covariates were included because of either a significant effect on the health outcome or based on rationale. Statistical significance was assumed at P < .05(two-sided).
Discussion
We assessed the association between two-year changes in physical activity levels and health outcomes in post-bariatric surgery patients. In addition to our previous work (
12), we revealed that only 38% of bariatric patients had substantially increased physical activity levels at 2 years post-surgery. These patients had greater weight loss and greater improvements in body composition, CRF and HRQoL compared to the patients who did not change or decreased their physical activity. Our findings underscore the importance of perioperative bariatric care programs to induce sustainable lifestyle improvements and improved habitual physical activity levels.
We found a positive relation between improvements in physical activity and weight loss following bariatric surgery. The magnitude of weight loss among the increased PA group (−43.0 ± 14.7 kg) was lower compared to findings reported in a previous study (−52.5 ± 15.4 kg) (
20). However, this study defined increased physical activity by reaching physical activity guidelines and therefore probably also reflects greater increases in PA compared to our study. We defined change in activity as the difference between pre- and post-surgery physical activity rather than reaching specific cut-off values (e.g. reaching physical activity guidelines), since small increases in physical activity can yield large health benefits (
21‐
23). Our findings confirm that small PA improvements are associated with better long-term outcomes, which could be highly motivating for post-bariatric patients to become more active. Furthermore, a meta-analysis on the effect of exercise interventions on post-bariatric weight loss also showed 1.94 kg additional weight loss in exercise groups compared to controls (
24). However, these well-controlled conditions of supervised training programs can often not be reproduced in unsupervised real-life conditions, and do therefore not reflect adoption of a more active lifestyle. Nevertheless, two-year changes in PA are of the highest importance, since patterns of post-bariatric weight loss suggest that improvements up to 1–1.5 years post-surgery are ascribed to the impact of the procedure, whereas preservation of improvements is related to changes in lifestyle (
25). This pattern is supported by our data, since peak weight loss occurred around 15 months, whereas weight loss ranges became substantially larger around 2 years (supplemental Fig. 1). The larger inter-individual variation is likely the result of substantial differences in lifestyle factors, such as PA. Therefore, further improvements (and possibly less weight regain) are expected when increased PA levels can be sustained. Taken together, our findings promote the value of post-bariatric care programs for achieving changes in lifestyle and improvements in PA levels.
Preservation of FFM is becoming a relevant topic in post-bariatric care, because of the essential role of FFM in several metabolic processes and health. In contrast to our hypothesis, no association between physical activity and absolute FFM (kg) was found. This might be explained by the strong association between FFM loss and weight loss, and the large variation in weight loss within our population (
26,
27). Although excessive loss of FFM tissue might be harmful for underlying metabolic processes, eventually restoration of a health body composition (i.e. a higher %FFM) is probably more important than preservation of absolute FFM (kg) (
28), because of its strong relation to several cardiovascular risk factors (
29). Our results suggest that increasing physical activity is related to greater improvement in %FFM. Therefore, increasing physical activity could have a positive impact on cardiometabolic health, by achieving a relatively larger muscle mass.
Another valuable finding of this study is the strong association between physical activity and CRF. CRF is an important predictor of mortality risk (
30) and it is known that people with morbid obesity suffer from a poor CRF (
31). Although all PA groups showed improvements in relative VO
2max (ml/kg/min) (i.e. relatively more ml of oxygen available per kilogram body weight), absolute VO
2max (L/min) was exclusively improved by maintained- or increased PA. suggesting an increase in maximal oxygen uptake and a gain in aerobic capacity during exercise. The decreased PA group showed a reduction in absolute VO
2max (L/min), suggesting that no real aerobic gain was obtained in this group. These findings highlight the essential role of physical activity in CRF, considering that weight loss on itself is not sufficient to improve physiological measures of CRF (
32). A large meta-analysis showed that even a small increase of 3.5 ml/kg/min of oxygen consumption in CRF (corresponding to e.g. 1-km/h higher walking speed) is already associated with a 15% and 13% reduction in all-cause and cardiovascular mortality, respectively (
33). Thus, the 3.9 ml/kg/min greater increase in CRF from our increased PA group reflects a substantial gain in health compared to the decreased PA group.
To the best of our knowledge, this is the first observational study that examined the relation between physical activity and HRQoL at more than 12 months post-bariatric surgery. Some previous studies assessed HRQoL and physical activity within the first year post-surgery (
20,
34), however a longer follow-up is needed to assess the effects of adopting a more active lifestyle. Our large cohort of patients with 2 years follow-up demonstrated that an increase in physical activity is related to greater improvement in both physical- and mental health. Especially the positive association with the mental components of HRQoL is interesting, since bariatric surgery on itself does typically have a greater impact on physical aspects compared to mental aspects (
35). Therefore, physical activity could potentially contribute to an improved mental well-being in these patients, next to the general physical benefits.
It is unclear why some patients are able to improve their habitual physical activity levels and others are not, despite the same post-bariatric care program. The present study showed substantial impact of patient characteristics on the ability to improve PA, but the increased PA group showed a higher prevalence of hypertension, sleep apnoea and back complaints. Furthermore, the increased PA group was significantly more inactive before surgery, considering that 43% walked or cycled less than 15 min per day and only 9% participated regularly in sports. Therefore it is easier to improve PA for these patients by engaging in sports activities or increasing walking time. Previous studies also suggest other barriers to exercise in post-bariatric populations, such as physical barriers and discomfort, self-presentational concerns, resources, time, environment, and intrinsic motivation (
36‐
38). It is likely that these factors played a role in our study as well, therefore focusing on these aspects might improve post-bariatric care.
One limitation of our study is the relatively short follow-up of two years, since this period is considered the very end of the honeymoon period at which failure rates start to rise. Therefore, greater impact of life style factors such as PA would be expected with longer follow up. Nevertheless, the follow-up of 2 years is still interesting because our population passed the tipping point of weight loss (
39) and patients entered the normalization phase in their personal environment. Further research into longer-term effect of bariatric surgery, physical activity and weight loss is warranted. Another limitation is the use of self-reported data for physical activity, because self-reported data can be influenced by social desirability and is less valid and reliable than objective measures (
40). However, by including longitudinal data, our study obtains more insight into relations between physical activity and outcome parameters than cross-sectional studies. Furthermore, we had a low compliance of 31.2% in CRF, which was predominantly caused by the limitations of the Åstrand test protocol, such as weight exceeding the bike limit, physical complaints (e.g. joint injuries), health issues (e.g. severe hypertension) and use of beta blockers. Nevertheless, compliance rate was not different across PA subgroups (~30% of each subgroup), suggesting no selection bias and leaving a sufficient sample size (
n = 1232) to determine the relation between change in physical activity and CRF.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.