Failure of weight loss is the bane of bariatric surgery. All bariatric surgeries work. They can all fail, and do.

The Roux-en-Y gastric bypass is considered the gold standard of bariatric surgery, the mark against which all other operations are compared. In the last few years there have been two papers looking at the long-term success of gastric bypass. The first, by Kelvin Higa, looked at his series from the 1990s. A total of 242 patients were operated on between 1998 and 1999 [2]. Office follow-up was 7 % at 10 years. Telephone follow-up was 19 % at 10 years. As a group, the available patients had an excess weight loss (EWL) of 57 %. It turned out that 86 (35 %) had one or more complications during follow-up, including internal hernia in 16 % and gastrojejunal stenosis in 4.9 %. In this group 65 (33 %) failed to achieve EWL greater than 50 %. Only 136 (51 %) had nutritional testing at least once after postoperative year 1. Of these 136 patients, only 24 (18 %) had remained nutritionally intact.

This is a bypass paper from a fabulous surgeon, with one of the biggest bypass experiences in the world. It reflects the reality, and the difficulty of maintaining long-term follow-up in bariatric patients—you will hardly ever see them again, they are often malnourished and they will often regain weight.

More recently, Obeid reviewed a similar cohort of patients from NYU [3]. He studied 328 gastric bypass patients done at NYU between 2000 and 2004. After 10 years, he had found 134, or 46 % of these patients. They had 59 % EWL. Almost exactly as with Higas’s group, 35 % failed to achieve 50 % EWL after 10 years. Furthermore, 20 % had a BMI greater than 40 kg/m², and if their starting BMI was greater than 50 kg/m², it was 39 %. Of these, 9 % had revisional surgery. With regard to complications requiring surgery, 12.8 % had internal hernia , 6 % small intestinal obstruction and 3 % incisional hernia . Most of these complications occurred at around 3 years after surgery. Once again, as in Higa’s study, 87 % had at least one major vitamin or mineral deficiency .

The similarity between these two papers is striking. Yes, bypass works very well for many patients, but about one-third fail, and success comes at considerable cost. However, in the big picture, it is a successful procedure. The fact that one-third do not achieve a sustained weight loss that we consider satisfactory does not prevent surgeons performing this procedure in huge numbers, as it helps many, many people.

It is hard to imagine a more successful bariatric procedure in terms of weight loss than biliopancreatic diversion, with or without duodenal switch , especially when using Scopinaro’s original 50 cm common channel. Yet it can fail, and does. Years ago I published a paper about placing lap bands over gastric pouches in failed BPDs [4]. Gagner similarly published on resleeving the pouch [5]. Back in the mid-1990s, when we were doing BPDs, it was almost inconceivable that we would have to do this with this most aggressive operation. But we did. All operations work and they can all fail. It was the metabolic consequences that staunched my enthusiasm for BPD, not the occasional failure. In recent times, BPD has had a second coming, as revisional procedure for failed gastric sleeves .

It is easy to blame the patient, their food choices, their lack of exercise, their failure to follow up, but as Buchwald recently wrote, as often as not the operation has failed them [1]. Our job is to continue to help them find another solution for their obesity, usually with another procedure, which leads us to the laparoscopic adjustable gastric band, or LAGB. The LAP-BAND® (Apollo Endosurgery, Austin, TX) was the first commercially available LAGB in the USA. The band was developed by Vern Vincent, then at Inamed, who expanded on the permanent band used by Kuzmak, making it adjustable and suitable for laparoscopic placement. The new band was first placed by Belachew in Belgium in 1993, closely followed by Favretti in Italy, and Himpens in France [6, 7]. Its appeal was obvious. Placement of a band involved little dissection, it was adjustable, it could easily be removed if it failed, and, in many cases, it worked very well. It appealed to many patients who simply would not agree to any of the more aggressive procedures that involved cutting and rerouting organs, as well as significant nutrient deficiencies [8].

The band became very popular in Europe, then Mexico, then Australia, after O’Brien, who had a long experience in gastric bypass, visited Belachew and Favretti and brought the technique back home. He taught me how to place the band, and most importantly, the philosophy of regular, frequent band adjustments , aimed at enhancing satiety, and most important, reducing the relentless hunger and the constant urge to eat, that bedevil obese patients.

As a morbidly obese man myself, with multiple comorbid conditions, and on 11 medications, it seemed a pipedream that I could be not hungry. I started doing bands in 1995, heard over and over that patients were less hungry, saw many of them lose a lot of weight. Eventually, in 1999, I had my own LAGB placed by Paul O’Brien. For me and many others, the pipedream came true. If a band is well-adjusted, patients are less hungry, have less urge to eat, eat less, and lose weight. If the band is kept adjusted, they do not gain weight. That is the essence of success with a band: adjust it correctly, and keep it there. Dixon stressed the importance of not over-tightening the band, of keeping the band and the patient in the “green zone,” not too loose and not too tight. He also showed clearly that the band reduced hunger [9, 10].

One of the confounding issues with the band is the variability of outcomes around the world. Some centers have great results, others in the same city not. There has been a great diminution in band placement over recent years, partly as a result of this variability. To me, having done bands for 20 years, and having had one for nearly 16 years, I fully understand its variability, even day to day. Before I discuss possible reasons for band failure, and need for further surgery it is important to present a few long-term studies of band outcomes, and assess the impact of this reduction in band placement.

In many facets of surgery, lesser is considered better if the outcomes are similar. Then, if the lesser procedure fails, one can move on to more complicated approaches. The classic example of this is cardiac disease, where there is usually a steady progression from statins to stents, to more stents, to CABG, to redo CABG. Likewise with knee and hip joint surgery, going from resurfacing to joint replacement, to repeat joint replacement if needed. These progressions are not considered failures, but rather continued steps in the treatment of a chronic disease. Obesity is a chronic disease. It seems logical to consider such a progression in our treatment. This is why the LAGB appealed to so many patients, and their doctors. As we have seen, the same applies to BPD, bypass, and most certainly to sleeve gastrectomy. Many of them will need revision. With the band, its initial safety and effectiveness made it a perfect first operation, providing the outcomes were equitable with the more complex procedures.

There is little doubt that the band is safe. In a 2012 study of UHC academic hospitals, Nguyen et al. reviewed 10,151 bands placed between 2007 and 2009 [11]. Mean length of stay was 1.2 days. Morbidity was 3.0 % and mortality only 0.03 %. Over a 3-year period band revision occurred in 0.76 % and explantation in 0.87 %. Gould, in a study of 32,509 bariatric patients between 2005 and 2007, of which 21 % were bands and 79 % a mix of open and laparoscopic bypasses, found mortality very low in both: 0.02 % for band and 0.09 % for bypass. However there was a significant difference in inpatient complications: 1.6 % for band and 4 % for bypass (p < 0.01) [12]. Saunders reviewed 1 year readmissions in a high volume bariatric center and showed readmission after a band of 12.7 % and 24.2 % for a bypass [13].

Buchwald was the first to show that the outcomes were equitable [14]. He did a meta-analysis of 22,094 patients in 136 studies, and found the band had 61.6 % EWL with 0.1 % mortality, compared to the bypass with 68 % EWL and 0.5 % mortality. The BPD had 70 % EWL, with 1 % mortality. The difference between 61 and 68 % EWL in an average patient with a BMI of 47 kg/m² is about 15 pounds. Is a 35 year old 5′ 4″, 300 pound woman who loses all her comorbidities, gets a better job and loves her new life really less a success if she ends up 185 pounds instead of 170?

These results were mid-term. As we all know, it is the long-term that matters. Data on bypasses over 10 years show EWL around 57 %. At NYU, Weichman reviewed 2909 patients with an average BMI of 45 kg/m², who had bands between 2000 and 2008 [15]. There was an operative mortality of 0.06 %. At 7 years they had maintained a 47 % EWL. O’Brien, in a series of 3227 bands between 1994 and 2011, had 714 patients at longer than 10-year follow-up [16]. There had been no perioperative mortality for primary placement or any revisions. He also found 47 % EWL in this group. Only 5.6 % had their bands removed.

The band is safe, and it often works. Why does it sometimes fail? The mechanism of weight loss after a band is two-pronged. The objective is to adjust the band to achieve loss of hunger and increased satiety. It needs the patient to attend regularly, and the surgeon or their team to make these visits available, then adjust the band as needed. Then, the patient has to eat slowly, slower than they ever thought. If either of these components is missing, the patient will regain weight.

Ever since O’Brien, Dixon, and I started teaching about the band, it has been all about the adjustments. Patients want the band and they want it to work. For it to work, it needs to be adjusted. It certainly does not fit the usual surgical model, where an operation is done, a follow-up visit or two are made, and the patient is discharged. We still see patients once a month for about 18 months then less frequently, but still regularly, up to 5 years, then yearly or as needed. I tell patients we are titrating their medicine, much like adjusting their blood pressure meds, perhaps needing to add another one if they need more. It becomes a rhythm of their life in the first couple of years, coming once a month for an adjustment.

On average, if they do 12 visits in the first year, they will get six adjustments. If they are losing weight, usually about 2 lbs a week, and they are not hungry, they do not need a fill. One cardinal sign is if the patient wakes up and craves breakfast. For some reason, well-adjusted band patients are never hungry in the morning. Furthermore, due to the diurnal functioning of the esophagus, they get full much quicker in the morning. Likewise, if they are snacking after dinner they are too loose.

So the first reason band patients fail at weight loss is that they either do not attend for follow-up, or do not get adjusted when they do. It is a lot of work, for both parties. Patients need to find the time to attend clinic. Surgeons need to work out dealing with the caseload in their office. As an example, at NYU we do about 1200 bariatric cases a year. Many are still bands. We do about 1000 band adjustments a month. The majority are in our main office, but many are done at satellite clinics, where we rent space on a weekly or monthly basis, around the periphery of New York. This makes it easier for patients to attend, so they do.

In the first year, patients need these regular adjustments, as they quite quickly adapt to the level of tension on the band. If they wait another month, they will be too loose, and they will regain weight. Seeing band patients every 3 months in the first 2 years after surgery is an exercise in futility. If they are seen, say they are still hungry and eating too much, yet are told to change their eating habits, rather than having their band adjusted, that too is a waste of time. What is more, they become discouraged, see little point returning, and do not. Then a few years later they turn up, and everyone agrees the band does not work. It was never given a chance to.

As a reflection of this, a study from the Netherlands reviewed 201 patients given bands between 1995 and 2003 [17, 18]. Of these, 193 were longer than 14 years. Nearly half, 46 %, still had their bands. They were only seen six times in the first 2 years, and once a year after that. It is surprising that any had done well. This is reflected in the poor weight loss compared to their patients who had gastric bypass. A similar study from Finland, of 60 patients at 14 years, with 100 % follow-up, showed that 52 % still had their bands, and they had 49 % EWL. The main difference between these two studies and many like them, and data from O’Brien in Australia and us at NYU is the incidence of band removal . I will discuss that later. The striking similarity is that in those patients who kept their bands, the weight loss in all these studies is the same, about 50 % EWL long-term.

At NYU, we have been referred 441 patients who had their band elsewhere, and were unhappy with their outcome [19]. Many had been told to have their bands removed, but preferred not to. Of these, 293 needed a band revision , and 26 patients had band removal and conversion to another procedure. The remaining 222 patients were treated by band adjustment , dietary counseling, and behavioral counseling about how to live with a band.

The most important points they were taught included the variability of band tightness due to stress, that it is always tighter in the morning, to eat smaller bites, and most important, to eat slowly, waiting between the bites. We also discussed that they would not become malnourished by eating such small meals, and that if you are not hungry, do not eat. If you are hungry, try hard to eat something healthy. At the end of the day, the band determines portion size and desire to eat, while the patient determines what they eat. If most of what they eat is healthy, they will do well.

When they presented, these patients had lost an average of 12.2 % EWL. After only 1 year, they had a further 24.5 % EWL after following this advice.

Once a patient has a band and it is adjusted, it is quite tight. You simply cannot eat at the rate you used to. You cannot live as if you do not have a band. You cannot eat at the same rate as other people at the table, ever. It takes about 20 s for food to go down the esophagus, into the pouch and through the band. If you eat quicker than that, the bites pile up above the band, stretching the esophagus. This causes a reflex to gag and regurgitate that is terribly unpleasant, and embarrassing in social settings. This is the second reason bands fail. Patients do not eat slowly enough.

I constantly stress this, from the very first consultation. Every visit, I make patients look at a clock and see how long 25 s really is. I advise them to get a 30-s timer and watch the sand fall. I tell them to put their knife and fork down, put the sandwich or the slice of pizza down between every bite. Let it go down. I have heard more nonsense about foods people cannot eat with a band than I ever thought possible. If you eat slowly, in most cases it is fine. If you do not, food gets stuck and you are in the bathroom bringing it up. As I tell every patient, I have had a band for all these years, and if I eat at the same rate as my slender wife, I puke. It is not negotiable. The problem is that many people do not like being told what to do, least of all by a thing. So they eat fast, vomit, and come in to have the band loosened, which renders them hungry, and able to eat easily, so they start regaining weight. Or they switch to soft food like ice cream that goes down easily, and gain weight. It is all about eating slowly. If I eat slowly, I eat less than my slender wife, I’m satisfied, and I keep the weight off.

This is all nonnegotiable. It will never change. It is why weight loss curves with a band are dead straight at about 50 % EWL for years and years. It never changes, as long as it remains adjusted.

One of the common catch-cries about the sleeve gastrectomy, which I hear every day, especially from patients who have had their bands removed to have a sleeve, is that it is so much easier to eat with a sleeve than with a band. Yes it is, especially after a couple of years, and that is why many of them are now dealing with rapid, total weight regain. If the band is kept adjusted, and patients eat slowly, they do not regain weight. If it is loosened due to the discomfort from eating too fast, they all regain weight.

The third cause of weight gain is a combination of the first two. Patients eat fast, food gets stuck, they loosen the band, they like how it feels, then they do not come back. They will regain every pound they have lost. If they do come back, and start again, they will re-lose the weight. More often, they feel the band has not worked, and prefer to convert to another procedure.

The next group of causes for failure of weight loss involve technical issues with the band. Of these, by far the most common is slipped band, often in combination with pouch dilatation or hiatal hernia . The main issue with a slipped band is reflux, which is usually severe, causing loss of sleep, inability to eat due to dysphagia , and at its worst, aspiration pneumonia. The diagnosis is easily made by an esophagram. The first step after diagnosis is to loosen the band, which leads to relief of symptoms, and immediate weight gain.

Over the last 20 years, the main efforts in the development of best technique for a lap band placement have been directed at reducing the incidence of slipped band. The first was the move to pars flaccida technique [20, 21]. Then in 2003 Dolan wrote of the importance of looking for and repairing hiatal hernias at the time of band placement [22]. In 2008, Gulkarov reviewed our experience at NYU, demonstrating a much-reduced need for reoperation if a hiatal hernia is repaired at the primary operation [23]. Between July 2001 and August of 2006, 1298 patients underwent a lap band and a further 520 patients underwent band with concurrent hiatal hernia repair. The mean initial weight and BMI were 128 kg (range 71.1–245.7 kg) and 45.4 kg/m² (range 28–75 kg/m²). Average follow-up for the band and band/HHR groups was 24.8 and 20.5 months, respectively. Rate of reoperation for HHR alone, or with band slip or concentric pouch dilatation, for band and band/HHR groups was 5.6 % and 1.7 %, respectively (p < 0.001). Total reoperation rate for slip, HHR and pouch dilatation was 7.9 % and 3.5 %, respectively (p < 0.001). There was no significant difference in rate of slip repair alone between the two groups: 2.3 % and 1.7 %, respectively (p < 0.44).

The importance of doing a hernia repair if one is seen, even a very small defect, became obvious to us. It is so important to reduce the chance of severe reflux which will require band loosening. The constant high pressure from the band will gradually dilate the pouch and blow open any weakness in the hiatus, eventually leading to a symptomatic hiatal hernia.

I believe the second cause of slip is excessively tightening the band. Loss of hunger and increased satiety is a wonderful thing for an obese patient. Not surprisingly, many feel that more is better. It is important to resist the entreaties to over-tighten the band. If patients are doing well, losing weight, and feeling well, do not tighten the band. If a band is too tight, it leads to pouch dilatation and to slipped band. The enemy of good is better.

The big decision facing a surgeon confronted with a slipped band is whether to fix it, or remove the band. I will discuss what to do upon removing the band later. This decision is of paramount importance. It is confounded by the fact that patients with a slip lose a lot of weight, love that fact, and are nervous that if they present for relief of their reflux, the band will be removed. As a result, they often present late, with a big slip. This can present technical difficulties during band revision.

The first step when facing a slip is to empty the band. I always warn the patient that they will be starving hungry within 24 h, and that they will regain a lot of weight, much of which is water. I then see them at 2 weeks and repeat the esophagram. Many times, the band will have returned to a completely normal position, and one can start re-tightening the band. If it is improved but not completely better, I repeat the process, often adding appetite suppression medications such as phentermine for the next 2 weeks. If it has not improved they need revision of their band. If the band is left empty, all patients will regain all their lost weight. The return of hunger, and no restriction of intake, is impossible to overcome.

With the surge of popularity of the sleeve gastrectomy, many surgeons opt to remove the band, and, pending insurance approval, do a sleeve gastrectomy at the time. Alternatively, they will remove the band and then enroll the patient in the required 6 months follow-up prior to insurance approval for a subsequent procedure. Much of this decision is predicated on the belief that band revision is technically challenging, and that patients will not be able to keep their weight off after revision.

Band revision surgery can indeed be challenging, largely due to the thick fibrous ring that forms around the band, especially posteriorly around the left crus of the diaphragm. Like all operations, it is a matter of step by step. After insufflation of the abdomen, I place a liver retractor and three 5 mm ports, using the old scars. If the patient had a single incision band placement, I gain access in the left subcostal area, and assess whether extra ports need to be placed, or it can be done through the periumbilical incision.

The first step is to divide all adhesions to the liver, to completely expose the hiatus and enable full liver retraction. I then start around the buckle of the band, using a diathermy hook to divide all the adhesions to the buckle. Once it is free, the band can be rotated easily, and also used as a retractor. Then I sharply divide the plane between the gastrogastric sutures over the band. The band is then pulled to the right to allow further division posteriorly. Once the gastrogastric suture plane has been divided, I fully mobilize the fundus off the left crus of the diaphragm, to visualize the gastroesophageal junction. I then assess whether there is a hiatal hernia, and if so, repair it anteriorly with figure-8 sutures of 0-polypropylene. I then unlock the band, slide it up to its correct position, relock it and redo gastro gastric sutures.

Yes, it can be difficult, but it is very worthwhile for many of these patients to be able to keep their bands, especially those who have done well.