Discussion
The main finding of this study was that weight loss was maintained for up to 5 years in our population of patients who underwent LAGB for the treatment of morbid obesity. On average, patients achieved nearly 50 % EWL by 2.5 years post-surgery, which remained steady in our population followed for up to 3 (
n = 719), 4 (
n = 304), and 5 years (
n = 74). The amount of EWL documented in the present study agrees closely with the findings from a review of 35 studies of diabetic patients who underwent LAGB, where weight loss was found to progress over the first 2 years post-surgery to reach a mean of 47 % EWL at 2 years. [
23]
The maintenance of weight loss with LAGB found in the present study is also in line with findings from a number of other publications. In an Austrian study of 276 patients, mean EWL was maintained at more than 65 % for 10 years following LAGB surgery. [
16] A French study of 140 patients showed an increase in EWL from 1 to 5 years following LAGB, for a mean of 46 % EWL at the latter time point. [
13] Similar results have been reported by other groups and in several meta-analyses, as summarized in Table
3, although positive long-term results are not universal. [
24,
25]
Table 3
Summary of % excess weight loss with LAGB in the published literature
| 148/127, ≥ 5 years | 46 | 7 years |
| 575/312, 2 years/66, ≥ 5 years | 40 | ≥5 years |
| 276/221 (estimated from 80 % follow up) | 64 | 10 years |
| 143/140 | 46 | 5 years |
| 151/82 | 43 | 13 years (median) |
| 442/135, 3 years/31, 5 years | 51 | 3 years |
60 | 5 years |
| Meta-analysis of 28 studies | 43 | 1 year (15 studies) |
50 | 2 years (12 studies) |
55 | 3 years (9 studies) |
| Meta-analysis of 18 studies; 4,456, 1 year/3,383, 2 years/640, 5 years | 42 | 1 year (11 studies) |
53 | 2 years (11 studies) |
55 | 5 years (5 studies) |
A number of studies have found that initial loss of excess weight is greater with gastric bypass than LAGB. [
3,
12,
14,
26,
27] However, an analysis of pooled data from 18 gastric bypass and 18 LAGB studies found that the total EWL over time was not different between the two procedures at the later follow-up time points (gastric bypass vs. LAGB: 62 vs. 55 % at 3 years, 58 vs. 55 % at 5 years, and 55 vs. 51 % at 7 years).[
3] This suggests that weight loss is more gradual with LAGB than gastric bypass but that it is just as durable and of similar magnitude (i.e., weight loss efficacy) over the long term. One study reported EWL at 7 years to be 58.6 % with gastric bypass and 46.3 % with LAGB in matched cohorts of >100 patients per group.[
25] The 48 % EWL with LAGB observed in this study is in line with the literature showing that patients lose a mean of 43 to 64 % of excess weight over the long term (Table
3). In the absence of randomization, it is not possible to conclude that the apparent differences between LAGB and gastric bypass reported in the matched cohort study [
25] are due to the procedures themselves (i.e., patient selection factors on which groups were not matched could have contributed to the differential weight loss).
Many studies have shown that resolution of obesity-related comorbidities depends on significant and sustained weight loss and excess weight loss [
28‐
30]. Results of a controlled study found that, among individuals who have had diabetes for less than 2 years, the disease remits in most patients who lose at least 10 % of their body weight following LAGB, whereas the disease does not remit in most patients who lose less than 10 % of their body weight following conventional therapy [
28]. Two recent prospective studies exploring mechanisms of type 2 diabetes remission have demonstrated that early improvements of insulin sensitivity and intracellular glucose disposition were secondary to caloric restriction shortly after surgery and from the amount of weight lost over time.[
31,
32] This suggests that the predominant effect of bariatric surgery on type 2 diabetes is due to weight loss, despite changes in gut hormones. Remission of diabetes was more likely to be observed in patients with a shorter history and better control of type 2 diabetes prior to bariatric surgery.[
31,
33,
34] A recent longitudinal study found that diabetes duration <4 years, body mass >35 kg/m
2, and fasting C-peptide >2.9 ng/mL were pre-operative factors predicting remission of diabetes at 1 year after gastric bypass.[
35] Taken together, these studies suggest that, in order to achieve remission of disease, surgical intervention should be considered at an early phase of diabetes in the obese diabetic patient.
Both LAGB and gastric bypass reduce body weight by approximately 20–40 % and excess body weight by approximately 50–75 % over the long term.[
3,
35,
36] Although weight loss at 1 to 2 years post-operatively is more rapid following Roux-en-Y gastric bypass (RYGB) or laparoscopic sleeve gastrectomy (LSG), excess weight loss for LAGB and LSG over time is similar (50–60 %) and somewhat lower than that reported for RYGB (60–75 %); however, morbidity at 1 year is lower for LAGB (5 %) compared to RYGB and sleeve gastrectomy (11–15 %).[
37] Given that gastric bypass, sleeve gastrectomy, and LAGB surgeries are all associated with clinically meaningful weight loss, the choice of bariatric surgery should involve careful consideration of both surgical and patient factors. Surgical factors include the expertise and experience of the treatment center, the risk of surgical complications, the complexity and reversibility of the procedure, and the availability of aftercare.[
37] In addition, the patient's weight loss goals, pre-existing comorbidities, willingness to comply with the required dietary and lifestyle changes, and the patient's preference are also important factors to consider when making the decision as to the choice of an optimal surgical procedure for each patient.[
37]
In our model, older age, male sex, and higher baseline weight were positively associated with post-LAGB WL and EWL over time. Higher baseline excess weight was negatively associated with EWL after surgery. Given the known limitations of %EWL as an outcome measure for weight loss (e.g., the heavier the patient, the smaller the %EWL) [
38], the effect of baseline WL (in the %WL model) provides a more clinically meaningful estimate. Older patients lost more weight than younger patients, and men lost a higher percentage of excess weight than women (although there was no difference between men and women in percentage weight loss).
The model used here is a novel approach to providing a more accurate assessment of post-bariatric surgery weight loss in the real-world clinical setting where the compliance of individual patients to recommended follow-up visits can be highly variable. The traditional approach to evaluating post-surgical weight loss parameters is based on simple arithmetic means, which frequently requires interpolation to “best-fit” the actual patient visits to pre-determined study time points and may omit patient data points that do not fit within specified “visit windows”, introducing additional sources of bias into the calculation of the population means. The advantages of the model used here are that it incorporates all data points for each patient, requires no data reduction or imputation, and uses the weight loss outcomes achieved by each individual patient to generate a statistically more accurate estimate for the mean weight loss outcome for the overall population.
Further studies are needed to determine whether the pre-operative patient characteristics used in this model can be used to provide a clinically relevant prediction of the post-operative weight loss in an individual patient over time. Until further data are available, these pre-operative factors may be of use to the clinician for counseling patients on expected outcomes based on the baseline characteristics of the individual patient.
Adverse events were relatively low in the overall patient population, with 238 of 2,815 patients (8.5 %) experiencing a total of 260 adverse events during a mean follow-up of 21 months and a follow-up of 5 years in 74 patients. The most frequent adverse event in our population was PPD (band slippage or pouch dilatation), which occurred in 118 patients (4.2 %). The only other adverse events that occurred in more than 1 % of patients were tubing/access port problems (
n = 35; 1.2 %) and band explantation (
n = 35; 1.2 %). The rates of adverse events observed in this study are on the low end of those reported in the literature,[
15,
18], although others have also reported similarly low rates [
17]. Adverse events may be influenced by surgical technique (e.g., pars flaccida, which was used exclusively in the present study, vs. perigastric) [
39], the type of band used, frequency of follow-up (but not necessarily adjustments), and surgeon experience [
11]. Nine (0.3 %) patients in this study population died over the 5 years encompassed by the analysis. In the eight patients for whom causes of death could be ascertained, none was believed to be related to LAGB treatment.
An additional consideration influencing the rate of long-term LAGB complications pertains to the differences in healthcare systems in various countries. In Canada, there may be a greater tendency to revise the band or port rather than explant the device because wait times for revision procedures such as gastric bypass surgery may be as long as 3 years. In contrast, revision surgery to remove the band and perform a second procedure may be more readily available in the USA and Europe. This may lead to higher rates of explantation or secondary procedures rather than band retention with revision. Other local and regional factors could influence the frequency of selected long-term complications and their management.
Additionally, the percentage of patients LFU in this study was low (19 %) considering its long-term nature. We recognize that patients who are lost to follow-up may not be captured in the reporting of adverse events, and it is possible that some patients who underwent band explantation did not report this to our clinic. However, the nature of the Canadian healthcare system and the limited access to revision surgery would likely encourage patients with adverse events to follow up with the clinic at which the band was implanted. LFU patients are not captured in the weight loss data beyond their last follow-up time point, and therefore we cannot make any assumptions as to whether weight loss was maintained, increased, or decreased over time. This limitation is common to all long-term studies of bariatric surgeries, many of which do not report the rates of LFU, thereby affecting the generalizability of the data presented.
Overall, the present study adds to a growing literature establishing the medium- to long-term stability of weight loss with LAGB. Notably, the 5-year weight loss experience in the outpatient setting reported here is consistent with other long-term studies demonstrating clinically meaningful and sustained weight loss outcomes with LAGB.[
3] Consequently, the maintenance of weight loss, combined with the low rates of adverse events and the ability to routinely perform the procedure in an outpatient setting, makes LAGB one of several effective surgical options for obese patients. As all bariatric surgery procedures require long-term behavioral and lifestyle changes by the patient, it is important that the clinician consider both the safety and efficacy of the surgery, along with patient preferences, to determine the most appropriate plan for the individual patient.