The Children’s Obesity Clinics Treatment protocol seemed effective in attaining a significant reduction in BMI SDS in severely obese children and adolescents. This reduction was achieved independent of baseline variables such as age, pubertal development, baseline obesity, socioeconomic class, co-morbidity, and place of referral and with a modest investment in manpower. As such our results are consistent with earlier reports evaluating the use of TCOCT [
17]. The degree of obesity at baseline was correlated with parental BMI, which is consistent with previous findings [
28]. We found no association with age in contrast to previous studies indicating that adolescents are less able to lose weight [
29,
30], although we did see a trend with patients below median age at baseline dropping out less than the older patients (p = 0.08). Marital disruption has previously been found to have a negative influence on a child’s weight [
31,
32]. In relation to this, we found that girls, but not boys living with a single parent reduced their weight less successfully compared to living with both parents. We achieved similar results as Holm
et al. [
17], with comparable time spent by health professionals per year and with lower retention rates. No baseline or longitudinal characteristics were significantly associated with an increased risk of dropout. Chronic care treatment is a long process, and while it is more efficient in reaching long-term success [
33,
34] than short-term treatments are [
35], it is also demanding in terms of attendance over time. A high attendance rate has earlier been shown to be an important predictor of success [
11], though we found no difference in treatment outcomes between patients who had been seen more frequently in the clinic than others. As this intervention is part of a public hospital setting, standard service was applied to patients and no extraordinary efforts were made to strengthen retention or attendance rates. Nor did we use any exclusion or selection criteria to sort out patients with severe co-morbidities or little motivation. Reinehr
et al. found a BMI SDS reduction of 0.36 in one year and 0.46 BMI SDS after four years. However, in the latter study only selected motivated families were included and efforts were made to achieve high retention rates, and baseline BMI SDS was 2.46 [
29]. Our study population consisted of severely obese patients (median baseline BMI SDS 3.0), and weight loss in this group might be more difficult than in less obese or overweight children [
36]. Reductions of more than 0.5 BMI SDS [
37,
38] and more than 0.25 BMI SDS [
39-
41] have previously been demonstrated to be a clinically relevant weight loss in terms of improvement in some cardiovascular risk factors and insulin resistance. In the present study BMI SDS was reduced by 0.40 BMI SDS in boys after 2 years of treatment. The proportion of patients achieving a weight loss greater than 0.25 BMI SDS increased over time, reaching 53% after 2 years, suggesting that remaining in treatment is beneficial. Even though we have not reported other measures of success, studies based on previous results from TCOCT protocol reported significant reductions in all fractions of cholesterols (3.8% lower per unit of BMI SDS) [
42], lower risk of hyperlipidemia (odds ratio = 0.37 per unit of BMI SDS) [
42], and lowered blood pressure [
43].
The present study makes a fair representation of the clinical reality in treating severely obese children and adolescents without any prior eligibility criteria and thus includes patients with both other diseases as well as obesity related complications. The reductions in BMI SDS were attained with 4.5 hours per patient per year spent by the clinical personnel. With more than 300 patients and a follow up of more than 2 years in 90 patients, this clinical study provides information about long-term treatment results for obesity. However, as per study design, a control group was not included. In principle, this means that the observed improvements in patient BMI SDS cannot be concluded to be the result of the treatment
per se, and could be due to other circumstances in the lives of the patients. However, as childhood obesity increases the risk of adulthood obesity [
5], a reduction in the degree of obesity in our patients while in treatment seems unlikely to be sporadic or stochastic in its origin. Even so, several other measures such as cholesterol, musclemass/fat free mass, blood pressure and psychological measurements would have provided more specific information about the treatment outcome. Furthermore, we did not follow patients after discharge or dropout, and thus have no data about outcomes after treatment intervention. With 25% dropping out after one year, we cannot neglect a potential dropout bias. Adolescents seem to be the most difficult of patients to target and treat [
13]. Other factors we did not measure could have biased treatment outcome, such as patient-family motivation and confidence.