A recent digest of top stories led with The Economist: back and shoulder surgeries are often no better than placebo, and imaging finds "problems" in people who have no pain. A strong topic, but that's not what caught me. In the comments, someone cited a popular statistic: Massachusetts has more MRI machines than all of Canada. The conversation quietly shifted to a different area—to what happens around a person heading into surgery.
Because there's a whole ritual built around surgery: a night without water, canceled herbal supplements, "just-in-case tests," a glass of sweet water a couple hours before anesthesia. Formally, this is medicine. In reality, some of these rules have never been tested, and some have been tested and produced inconvenient results that stopped no one.
So the question was: how much of the preoperative ritual is evidence, and how much is habit and institutional insurance?
By WHO estimates, more than 300 million surgeries are performed worldwide each year, and in wealthy countries, people are more likely to end up on an operating table than behind a wheel: in England, about 60% of people undergo at least one surgery in their lifetime. For a long time, checking whether a specific surgery was necessary was nearly impossible—truly blinded studies in surgery were almost never done. The situation changed sharply: the number of surgical randomized trials funded by Britain's NIHR grew from 34 in 2011 to 188 in 2023, with similar programs appearing in Australia, Canada, and across Europe.
The results were unpleasant. Removing an inflamed appendix is no better than a course of antibiotics for most patients. Spinal fusions and rotator cuff repairs produce effects indistinguishable from placebo or standard physical therapy. A British prostate cancer study found no impact of surgery on mortality after 15 years, but did find more sexual and urological problems. A Finnish study of a common knee operation found that after ten years, those who had surgery were worse off than those who received a sham procedure—anesthesia and a superficial incision without intervention.
Then it gets more interesting, because surgery is also a powerful suggestion machine. In 2022, authors of a paper in JAMA Network Open gathered data from one hundred surgical trials covering 32 interventions and calculated: on average, two-thirds of the improvement a patient feels after surgery is explained by natural healing or placebo effect. The actual procedure accounts for about a third. American orthopedist Seth Leopold puts it bluntly: surgeons are confident in their effectiveness because they "administer a powerful placebo to the patient."
The classic case is Bruce Moseley's 2002 study on 165 patients. Some received real knee arthroscopy, some received a sham: anesthesia, an incision, and the surgeon simply splashed water in a tray so the sound of "flushing" the joint could be heard. No difference in pain was found. The sound of running water was part of the study design precisely because the team assumed: the unconscious also hears what happens on the table.
Now the second layer—what a person does before surgery. And here the main surprise is that the guidelines themselves have never been tested in a controlled study.
The point of preoperative fasting is to reduce the risk of aspiration of stomach contents into the airways. This event is rare: in elective surgery about 2–3 cases per 10,000, mortality around 0.1 per 10,000. Randomized studies of "fasting versus aspiration" have not been conducted and won't be—too rare an outcome for such numbers to prove anything. An international consensus document published in 2026 in the journal Anaesthesia says this explicitly: previous recommendations relied on expert opinion and extrapolation of physiological data, not on outcomes.
What actually happens in clinics? Median actual fasting times are 14–17 hours without food and 9–12 hours without liquids, despite recommendations of 6 hours for solid food and 2 hours for clear liquids. In a survey of 971 anesthesiologists, 85% claimed their advice matched guidelines, yet half forbade patients from drinking after midnight. Both figures sit next to each other in one paragraph and illustrate the gap between declaration and practice quite well.
Fasting is not a neutral pause. The document lists consequences: reduced well-being, impaired glucose metabolism and enhanced inflammatory response, delayed return of peristalsis, loss of muscle strength. People arrive for surgery starving for a day—and this is considered a normal state.
The good news is that things have moved here. Modern consensus recommends solid food 6 hours before, non-clear liquids (milk, meal replacements) also at 6 hours, and clear liquids—water, tea, coffee with a small amount of milk, juices, lemonade—up to 2 hours before anesthesia. And then comes the most radical part: hospitals can allow drinking right up until the moment the patient is called to the operating room.
The approach is known as "Sip Til Send," supported by the Centre for Perioperative Care in Britain and the New South Wales government, already standard in many British clinics, and in the Australasian network there are more than sixty such hospitals. German and Dutch hospitals give patients fasting cards. A separate consensus point, with 92% agreement, covers candy and chewing gum—they can be consumed right up until transfer to the operating room.
There's a detail in this ritual that any patient at a large clinic knows: two hours before surgery, they bring a clear glass of sweet water—a "carbohydrate load." The logic sounds reasonable: surgical trauma causes insulin resistance, and a gulp of sugar supposedly prepares the body for this shock in advance. ERAS protocols, "enhanced recovery after surgery," recommend it strongly—yet, as authors of a 2025 systematic review note, the quality of evidence behind this recommendation remains low or moderate.
What the math shows. A 2022 Bayesian network meta-analysis gathered 4,936 patients and confirmed: a low dose of carbohydrates before surgery reduces postoperative insulin resistance, but after filtering out studies with high risk of systematic error, the effect shrinks from −3.25 to −1.29. There too—a conclusion that clinical guidelines don't like to repeat: there is no evidence that carbohydrate loading relieves patient discomfort. And another result that makes you uneasy: compared to regular fasting, low-dose carbohydrates were associated with postoperative infections, risk ratio 0.42.
A 2025 meta-analysis in Scientific Reports adds daily specifics: the carbohydrate group had hospital stays 0.56 days shorter. Half a day—with colossal heterogeneity between studies. The authors explicitly note that definitions of "length of hospital stay" diverge across papers, and the length itself depends on organizational factors like admission timing and discharge hour. I honestly don't know what to do with this half-day: it looks like a trace left by a very different set of very different hospitals.
Next—what the patient often stays silent about. A review in Baylor University Medical Center Proceedings: about a third of adult Americans take herbal supplements, and a significant portion don't tell their doctor. Supplement regulation works differently than for drugs, so they're simply absent from the patient's chart—but they do have pharmacology.
The breakdown by bleeding turned out unexpectedly selective. The only two supplements with a strong link to surgical bleeding independent of anticoagulant use are garlic and hawthorn. Echinacea, cordyceps, and aloe have a weak association. Ginkgo, turmeric, melatonin, chondroitin with glucosamine, blueberry, chamomile, fenugreek, milk thistle, and mint are dangerous, but only in people already taking anticoagulants: they have no evidence of bleeding on their own. Fish oil, ginseng, and saw palmetto didn't make it into the analysis at all—no association detected. And for St. John's wort, ginger, ginkgo, and cranberry, the data directly contradict each other.
Here I have an honest conflict with reality: hospital pamphlets list ginkgo in bleeding risk lists, while systematic review says there's no evidence without anticoagulants. I won't decide who's right—but I'll note that in warning lists, the logic of "better safe than sorry" works right up until the moment when a patient cancels a necessary medication out of fear based on nothing.
And the scale of silence is concrete. In a survey of 300 patients before surgery, 48.4% reported using folk medicine before surgery; in global studies the range is 12% to 69%. Non-disclosure is described by doctors in another study as a problem "related to the doctor themselves": the patient doesn't tell because no one asked in a way that sounded serious. The general recommendation repeats across different papers: non-essential supplements should be discontinued about two weeks before intervention.
The most revealing story in all of this—preoperative tests. Here everything has already been figured out, published, and approved in clinical guidelines.
A Cochrane review of routine medical testing before cataract surgery collected three randomized trials, 21,531 surgeries, and 707 significant adverse events. In the group that underwent full preoperative examination, 353 events occurred; in the group without tests or with minimal ones—354. Odds ratio for intraoperative complications 1.02, for postoperative 0.96, meaning nothing. But the cost in one study turned out 2.55 times higher. The authors' conclusion sounds like a verdict: routine preoperative testing does not improve the safety of cataract surgery. In the Choosing Wisely campaign, this recommendation stands as a separate item: don't do routine tests before low-risk procedures.
Notice the date. We're not talking about a fresh discovery, but a long-established fact, against which in real clinics there are still lines at the laboratory.
The answer to this question in surgery is unpleasant: because you have to retrain people who don't doubt. Bristol surgeon Jane Blazeby speaks about this with an intonation difficult to convey in writing: "You have to remember that surgeons have no uncertainty. They always know best." Her formula for negotiating with colleagues about randomized trials—teach them to "be confidently uncertain."
Then boring patterns take over. The probability that a surgeon will stop doing a procedure after a study showing its ineffectiveness drops with their experience. Surgeons who have been operating for years overestimate the effect: mainly those who got better come back for follow-ups, while those who weren't helped don't want to disappoint the doctor. For one common knee operation, a surgeon's prediction of "will improve or not" was at the level of a coin flip.
And most importantly—the environment. In countries where surgeons receive a salary, abandoning an ineffective operation happens quickly; where they're paid per operation, everything persists. In Britain, removal of a bone spur on the shoulder dropped from 28,000 cases in 2016–17 to 5,720 in 2019–20 after publication of results. In the US, demand for the same procedure remained.
A separate force—the patient: scans find tears natural for age, like wrinkles, and the promise of a quick fix wins. Some surgeons fear a bad online review, insurers fear losing a client and continue covering procedures that don't work. In a 2020 review, only 1% of studies on common surgeries for chronic musculoskeletal pain actually compared surgery with its absence.
The timeline is also worse than it seems. Analysis of clinical guideline recommendations in CMAJ showed that after three years about one in five becomes outdated: 92% are valid after a year, 77.8% after four. Waiting more than three years before revision is already too long.
If you add it all together, an almost comical picture emerges: the most proven part of the preoperative ritual is permission to drink water and suck on candy. Everything else is either untested or tested and ignored. The patient has exactly three levers, and all are boring: tell about all supplements, including "harmless" ones, ask which specific tests are needed and why, and find out what will happen if the surgery isn't done.
The wrinkles on the MRI, by the way, aren't going anywhere.