Bone Loss, Trace Minerals and Anemia: Three New Papers on Nutrition After Metabolic Surgery

The NIDDK’s plain-language page on the side effects of weight-loss surgery draws a line between the immediate and the late. Bleeding, infection, leaks where the stomach or intestine is stapled, and blood clots belong to the first group. The second is quieter: the digestive tract may no longer absorb enough nutrients, and if the prescribed vitamins and minerals are not taken, the result can be anemia or osteoporosis — conditions that announce themselves years after the operation rather than weeks.

Earlier work tracked in this archive has mapped the early complications, including the bleeding and dehydration-driven readmissions that cluster in the first month. Three papers published in Obesity Surgery over the past several days take the later horizon instead: a nine-year bone-density follow-up at a single centre in Warsaw, a multicentre audit of what actually gets measured across England, and a machine-learning model built to predict anemia twelve months out. They address the same worry at three different levels of evidence, which is the first thing worth keeping straight when reading them side by side.

Nine years of bone density in a small prospective cohort

Kędzierska and colleagues followed consecutive patients who had sleeve gastrectomy at one centre between 2010 and 2016, scanning the lumbar spine and femoral neck with dual-energy X-ray absorptiometry at baseline and again at follow-up. Seventy-two people completed both scans — 62 women and 10 men, mean age 43.6 — with a median of nine years between them.

Weight fell from a median of 112.8 kg to 91.5 kg, and BMI from 41.2 to 34.3 kg/m². At the femoral neck, bone mineral density dropped by a median of 0.10 g/cm², while the T-score fell 0.90 and the Z-score 0.60; all three changes were statistically significant. The lumbar spine told a flatter story. There, BMD and T-score changes did not reach significance, and only the Z-score moved, by 0.10.

The most interesting result may be a non-result. Changes in bone density showed no significant correlation with baseline values, with how much weight was lost, or with shifts in fat mass or lean mass. The authors conclude that bone mineral density “may be significantly decreased at the femoral neck” nearly a decade after surgery, and call for further prospective work to separate the mechanical, hormonal and metabolic mechanisms involved. It is a prospective cohort, but a small, single-centre one with no non-surgical comparison group — a signal about a site-specific change rather than a demonstration of cause.

What an audit of thirteen centres found

If bone loss raises the question of long-term follow-up, the second paper asks a blunter one: is anyone measuring the things that are supposed to be measured? British Obesity and Metabolic Specialist Society guidance calls for routine micronutrient monitoring and supplementation after surgery. Holt and colleagues, through the society’s BariDIET research group, gathered site-level data from thirteen English bariatric centres on adults who had primary surgery in 2021 — preoperative and three-, six- and twelve-month micronutrient levels, whether a clinician ordered the tests, and whether patients reported taking their supplements.

Testing was uneven early on. A clinician requested blood tests for 33.7% of the 669 patients at three months, 60.0% at six months and 88.4% at twelve months. Among those who were tested, low vitamin D and folate were less common at twelve months than before surgery. Zinc and selenium lows showed up in their place, and low zinc was more frequent after sleeve gastrectomy than after gastric bypass at every time point — 16.7% versus 11.5% at three months, 27.6% versus 10.0% at six months, and 32.2% versus 26.0% at twelve months.

The authors are notably careful with those percentages. Trace-mineral data were sparse and testing was selective, so the prevalence estimates rest on wide confidence intervals and a population that may not represent everyone operated on; roughly two-thirds of patients with recorded data said they took supplements as advised. The paper’s broader conclusion reaches furthest: inconsistent monitoring and substantial missing postoperative data suggest the nutritional guidelines are not being fully implemented in practice — and that implementation gap is itself what limits real-world evidence on deficiency after surgery.

Predicting anemia a year ahead

The third study inverts the problem. Instead of measuring what happens, Nie and colleagues tried to forecast it. Using records from patients aged 16 to 65 who had sleeve gastrectomy or Roux-en-Y gastric bypass, they built models from 45 preoperative variables in 511 patients at one hospital and validated them externally in 300 patients at two others, comparing logistic regression, random forest, support vector machines and gradient boosting. Anemia at one year affected 39.6% of the training cohort, 39.2% of the internal test cohort and 19.0% of the external validation cohort — a reminder of how much case mix can move a headline number.

A random forest reduced to seven predictors performed best, reaching an area under the curve of 0.827 in internal testing and 0.823 externally, and it outperformed blinded predictions from ten clinicians of varying experience, including on cases the clinicians themselves rated difficult. The authors present it as a possible aid to individualised monitoring and nutritional care. It is worth stating plainly that a prediction model is not a care pathway: this one was developed and validated on retrospective records, and whether acting on its warnings actually reduces anemia has not been tested.

What the three put together

Read side by side, the papers sketch an obligation that outlasts the operation. The bone findings rest on 72 people at a single centre; the deficiency rates rest on whoever happened to be tested; the model is a statistical instrument awaiting prospective use. None of them argues for or against surgery for any individual. A prospective cohort without a comparison group, a retrospective multicentre audit and a model-development study are not the kinds of evidence that settle a personal decision, and this archive — which is independent of NIH and NIDDK — does not offer advice of that kind.

What they do suggest is narrower and more useful: the nutritional and skeletal consequences of metabolic surgery are measurable, unevenly measured, and worth following with the same discipline as the weight curve. The LABS study, funded by NIDDK, was built on exactly that premise — the project’s longitudinal cohort and its substudies were designed to capture outcomes that only appear late. The open question these three papers leave is whether routine monitoring will catch up with what the guidelines already ask for.