What the anesthesia team monitors during sedation

Monitoring exists to catch a change while it is still small and still silent. Almost every anesthetic problem announces itself on an instrument well before it becomes visible to anyone standing in the room, and by the time it is visible the margin for a calm, unhurried correction has already been spent. Looking fine is not evidence of anything. It is the interval during which the instruments are the only things in the room that know.

Consider what a person can actually see. Colour is one signal: lips and fingertips turn dusky when oxygen falls. But that change does not become reliably visible until arterial saturation has dropped into the low eighties, and it is harder to see under operatory lighting, behind a drape, in a patient who is anaemic, and in patients with darker skin. By the time a room notices colour, an instrument has been describing the same event for a minute or more.

Chest movement is the other signal, and it is the more deceptive of the two, because a chest can move vigorously against an airway that is completely closed. Effort continues; the abdomen and chest rock against each other; nothing enters the lungs. From across the room that looks like breathing. On a capnograph it looks like a flat line, on the first breath, because there is no breath to show.

That gap between what an instrument knows and what a person can observe is the entire argument for continuous monitoring. It is not a formality attached to a routine procedure and it is not there to make an office look equipped. It is there because the observable signs of an anesthetic problem are, without exception, late signs.

There are five continuous or near-continuous streams and one intermittent one. They are not redundant. Each measures a different physiological system, each fails in a different way, and each has a characteristic delay between the moment something goes wrong and the moment the display changes. Knowing which is fast and which is slow is the difference between reading a monitor and being reassured by one.

Read that table as a single argument rather than six entries. No instrument on it is sufficient alone, and the fastest one is not the one most patients have heard of.

The clip on your finger is a pulse oximeter. It passes red and infrared light through the fingertip and calculates, from how much of each wavelength is absorbed by the blood that pulses through with every heartbeat, what percentage of your hemoglobin is carrying oxygen. That percentage is your oxygen saturation. It also produces a second reading most patients never notice: a waveform confirming that a pulse is genuinely arriving at your hand rather than that a number has been invented.

The number is useful and it is late, for three separate reasons that stack on top of each other.

The first is chemistry. The relationship between the oxygen tension in your blood and the saturation the probe reports is a curve that is almost flat at the upper end. Oxygen tension can fall from a normal value of around 100 down to about 60 while saturation drifts only from 100 per cent to roughly 90. A large, real, ongoing deterioration produces almost no visible movement on the display until it reaches the steep part of that curve, at which point the number falls quickly.

The second is plumbing. Blood that has just been oxygenated in the lungs takes time to reach a fingertip. Twenty to thirty seconds is typical, and it lengthens considerably when hands are cold or circulation is constricted, which is common in an air-conditioned operatory. A probe on an ear or a forehead reports the same event sooner, because the distance is shorter.

The third is the oxygen itself. A patient breathing supplemental oxygen through a nasal cannula has filled their lungs with a reservoir far richer than room air. That reservoir is protective, and it is precisely what delays the alarm: a patient who has stopped breathing can hold a saturation in the high nineties for minutes while carbon dioxide climbs and the airway stays shut. Giving oxygen makes the pulse oximeter a later warning, not an earlier one. That is a good trade, and it is only a good trade if something else is watching ventilation.

The readings that lie

Pulse oximetry is also the monitor most easily fooled. Dark nail varnish, particularly blue, green and black, absorbs light in the same range the probe uses and drags the reading down or stops it working. Cold fingers, shivering, movement and low blood pressure all weaken the pulse the device needs to find. Injected dyes interfere. And carbon monoxide is read as though it were oxygen, because carboxyhemoglobin absorbs light almost identically to oxygenated hemoglobin at the red wavelength, so a heavy smoker who smoked that morning can show a falsely comfortable number.

There is one further limitation worth stating plainly because it is a documented device problem rather than a theoretical one. A 2020 analysis published in the New England Journal of Medicine found occult low oxygen, meaning a saturation reading of 92 per cent or above alongside an arterial measurement below 88 per cent, roughly three times as often in Black patients as in white patients. The FDA has since issued a safety communication and draft guidance on how these devices are tested across skin tones. The clinical response is not to distrust the monitor. It is to treat saturation as one input among several, and to lean on the reading that measures breathing directly.

Capnography measures the carbon dioxide in every breath you exhale and displays it two ways: as a number, the end-tidal value taken at the very end of exhalation, and as a waveform showing the shape of each individual breath. Because carbon dioxide only reaches the sensor when gas is physically moving out of your lungs, capnography measures whether you are ventilating rather than inferring it from something downstream. That single property is why it warns earlier than anything else in the room.

A normal end-tidal value sits at roughly 35 to 45, and a normal waveform is close to a square: a sharp rise as exhalation begins, a flat plateau while alveolar gas comes out, a sharp fall as the next breath is drawn in. The shape carries more information than the number, and an experienced monitor reads the shape.

• The waveform disappears entirely: no gas is moving. This is apnoea or complete obstruction, and it is visible within one breath cycle rather than one minute.
• The upstroke becomes sloped and rounded instead of sharp: gas is leaving under resistance, which points at partial obstruction, a narrowed airway or bronchospasm.
• The number climbs steadily while the waveform stays normal: breathing continues but too slowly or too shallowly, which is the classic pattern of an anaesthetic that has gone one level deeper than intended.
• The number falls sharply with a normal-looking waveform: gas is reaching the lungs but blood is not, which is a circulation problem rather than a breathing one.
• The number rises steadily despite adequate ventilation: an uncommon pattern that is investigated urgently, because it is one of the earliest signs of a rare inherited reaction to certain anaesthetic agents.

The American Society of Anesthesiologists made continuous capnography a basic monitoring standard for moderate and deep sedation and general anesthesia in 2011, and Florida's anesthesia rules require exhaled carbon dioxide measurement at the depths that need a permit. The reason both did so is the failure mode above: airway obstruction under sedation is the common event, and it is the one pulse oximetry reports last.

The caveat specific to oral surgery

In an office the sample is usually drawn through a small tube in a nasal cannula. Third molar surgery is performed through an open mouth with high-volume suction running a few centimetres away, so a patient breathing through their mouth can exhale most of their carbon dioxide past the sampling port. The number then under-reads, sometimes considerably. Sampling cannulas designed for this setting carry an oral scoop as well as nasal prongs for exactly that reason.

This is worth understanding because it changes what the number means rather than making it useless. A low end-tidal value in a patient whose mouth is open may be dilution rather than hyperventilation. But the presence, rhythm and shape of a waveform still prove that air is moving, and that is the question capnography exists to answer.

Some oral surgery teams also use a precordial or pretracheal stethoscope: a small weighted disc taped over the chest or the neck, feeding continuously to an earpiece. It is old, cheap technology and it is genuinely fast, because the sound of air moving arrives at the same instant the air does. Nothing about the electronics has made it obsolete.

The adhesive dots on your chest and shoulders feed a continuous electrocardiogram. What it displays is electrical activity: how fast the heart is being told to beat and in what rhythm. Lead II is the standard rhythm view because the small P wave that marks the top chamber contracting is clearest there, and its disappearance or irregularity is the earliest electrical sign of several problems.

During third molar surgery the ECG earns its place for one reason more than any other. Manipulating tissue supplied by the trigeminal nerve, which is exactly what lower third molar surgery involves, can trigger a reflex slowing of the heart. It arrives suddenly, it is usually brief, and it is resolved by pausing the stimulus and, if needed, giving a drug. On an ECG it is unmistakable and instant. Without one it is invisible until the pulse becomes hard to find.

A rate that climbs has a longer list of causes: an anaesthetic that has become too light, pain reaching a patient who cannot report it, the adrenaline in local anaesthetic solution, anxiety at induction, or a drug effect. Which of those it is gets decided by looking at the other readings, which is the entire reason more than one thing is measured.

One honest limitation. A three-lead display in an office is a rhythm monitor. It is not a reliable detector of reduced blood supply to the heart muscle, which is what a five-lead configuration with a chest lead is set up to catch and what a full twelve-lead recording is required to characterise. If your history makes that a live question, the answer is a different assessment before the day, and sometimes a different setting, rather than a different display in the same room.

Blood pressure is the only reading in the standard set that is a snapshot rather than a stream. An automated cuff inflates on a cycle, commonly every five minutes and more often during induction or when something has changed, and reports a single value from that inflation. Everything about how it is interpreted follows from that: a fall in pressure can be almost a full cycle old before it appears, so it is read as a trend against your own baseline rather than as a live signal.

A modest fall after induction is expected rather than alarming. The agents used to produce and maintain anesthesia relax blood vessels and reduce the drive that keeps pressure up, and a drop of ten to twenty per cent from baseline is a routine, anticipated consequence. A steep or continuing fall is a different matter and is treated actively, with fluid, with position, and by reducing the depth of anesthesia.

Two practical details determine whether the number is worth anything. The cuff has to be the right size for your arm, because one that is too small over-reads and one that is too large under-reads, and the error is not trivial. And it goes on the arm opposite the cannula, so that inflating it does not interrupt the infusion or squeeze a drug into you in a bolus.

Why temperature is on the list at all

For a healthy adult having four third molars removed in a climate-controlled room over an hour, temperature rarely changes enough to matter, and it is generally measured when a change is anticipated, suspected or intended rather than continuously throughout. Its presence on the equipment list is about one specific event: a rare inherited reaction to certain anaesthetic agents, in which the body's metabolism accelerates uncontrollably.

The reason that matters in an article about monitoring is that the temperature rise is a late sign of it. The earliest sign is a steadily climbing end-tidal carbon dioxide in a patient who is being ventilated adequately, sometimes with unexplained muscle rigidity and a rising heart rate alongside it. The thermometer confirms; the capnograph suspects first. If the agents capable of triggering it are used, the antidote has to be stocked, and which agents a practice uses is a fair thing to ask.

A family history is the only warning anyone gets in advance, and it is inherited. If a blood relative has ever had a severe or unexplained reaction under anesthesia, that is one of the few pieces of information a patient can supply that no examination will reveal.

Depth is assessed by the same thing that defines it: how you respond. The four levels of sedation and anesthesia are written in terms of responsiveness, to your name, to a firm voice, to a physical stimulus, or to nothing at all, so testing your response is not a crude stand-in for a measurement. It is the measurement. The monitor speaks to you at intervals for that reason, and the answer, or the absence of one, is recorded.

Under general anesthesia there is no response to elicit, so depth is read from physiology instead: heart rate and blood pressure rising when the surgeon starts, movement, tearing, changes in breathing pattern in a patient breathing for themselves. These signs are cruder than most patients assume and they are what the team is actually working from.

The reason depth is monitored at all is that anesthesia is a continuum, not a set of discrete states. A dose calculated to produce moderate sedation carries some patients further than intended, which is why anyone administering it must be trained and equipped to rescue a patient one level deeper than the one they were aiming for. Watching the depth is how the team knows a patient has drifted before the airway tells them.

Brain-activity monitors, honestly

Processed electroencephalogram monitors exist. A sensor strip on the forehead converts brain electrical activity into a single index, with a range around 40 to 60 targeted for general anesthesia. They are used in some hospital anesthesia and they are worth understanding accurately rather than being impressed by.

The evidence that they reduce the chance of awareness is mixed. Large randomised trials comparing them against careful monitoring of anaesthetic agent concentration did not show the advantage that was expected, the index is affected by muscle activity and by which drugs are used, and it is not required in a Florida dental office. For a forty to ninety minute third molar case they are not standard, and a practice that does not use one is not cutting a corner. Any office claiming that a brain monitor makes anesthesia safe is overstating what the trials found.

What is required is a written anesthesia record kept at timed intervals through the case and into recovery. That record is not paperwork for its own sake. A number written down every few minutes creates a trend, and a trend is the form in which a slow drift becomes visible. A single reading in isolation almost never is.

The figures below are typical values used as a general frame. Your own targets are set against your baseline, your history and the plan for your case, and the surgeon's written parameters govern over anything general. They are given here because patients are routinely shown a screen full of numbers and told nothing about what any of them mean.

Notice what the right-hand column has in common. Almost every entry begins with a physical action taken by a person, not with a drug. Most anesthetic problems in a healthy patient are airway problems, and most airway problems are answered by a hand repositioning a jaw within seconds of the alarm.

Florida requires at least three trained individuals present at the chair for the whole of a general anesthesia or deep sedation case, under Rule 64B5-14.003: the operating dentist, an assistant, and a person whose sole responsibility is monitoring you. The third role is the substance of the rule rather than a detail of it. Monitoring is not a task shared with retraction, suction or instrument handling. It is one person's entire job for the length of the case.

The reason that separation is written into a rule rather than left to good practice is that divided attention fails in a specific and predictable way. A surgeon operating on a lower third molar is looking into an illuminated field a couple of centimetres across, through a mirror, with suction running and a handpiece turning. That is where their attention has to be for the work to be done properly. Asking the same person to notice a waveform changing shape is asking them to do the surgery worse or the monitoring worse, and in practice it is the monitoring that gives.

The delays described throughout this article are the other half of the argument. If capnography warns within one breath and the pulse oximeter warns a minute later, the value of the earlier warning is entirely determined by whether anyone was looking at it when it arrived. An unwatched monitor converts a fast warning into a slow one.

What that person is actually doing

• Watching the displays continuously, and specifically watching the capnograph waveform rather than only the numbers beside it.
• Keeping the anesthesia record at timed intervals, which is what turns individual readings into a visible trend.
• Maintaining the airway physically when needed, which for most events means a jaw thrust or a head reposition long before it means anything else.
• Responding to alarms first, so that the surgeon is told rather than interrupted, and the operative field is closed down in an orderly way if the case has to pause.
• Watching you rather than the clock during recovery, because discharge is decided against criteria and not against elapsed time.

The permit sits underneath all of this. Florida issues anesthesia permits by depth, and a practice may work only at the depth it holds. This practice holds a general anesthesia permit, and the equipment list, the staffing floor, the emergency drug set and the periodic drills are conditions attached to that permit rather than choices made around it. A Board consultant inspects the room and observes the team work through emergency scenarios before it issues.

One consequence patients notice: the permit holder stays with you from the start of the anaesthetic until you are discharged, which means anesthesia cases cannot be run side by side and are scheduled less densely than ordinary appointments. That is a scheduling cost the rule imposes deliberately.

Everything above describes detection. None of it describes prevention, and the distinction matters more than any individual reading, because a room full of equipment is easy to mistake for a margin of safety it does not by itself provide.

A monitor detects; it does not treat. The value of a capnograph is entirely contained in what a trained person does in the ten seconds after the waveform flattens. An office with the full equipment list and a team that has not rehearsed together is a worse proposition than the equipment suggests, which is why Florida requires drills and why asking when the last one was is a reasonable question.

Alarms are also frequently wrong. A cold finger, a shifted probe, a moment of movement, a kinked sampling line: all of these produce alerts that mean nothing. The dangerous adaptation is a team that has learned to silence a particular alarm because it usually cries wolf. Good monitoring practice treats every alarm as real until it has been shown otherwise, which is tedious and is the point.

Most importantly, the decisions that matter most were made before the day. Patient selection, a direct airway examination, an honest medication list, a fasting window that was actually observed, and a clear-eyed judgement about whether an office is the right setting at all. No amount of instrumentation converts an unsuitable office case into a suitable one. If your history points toward a hospital, monitoring in an office does not close that gap, and being referred onward is not a refusal.

It is also worth saying that none of this applies to a great many third molar extractions. A straightforward erupted upper third molar is routinely removed under local anaesthetic alone, with no fasting, no escort, no cannula and none of the monitoring described here, and for a patient who tolerates dental treatment comfortably that is frequently the appropriate choice rather than a lesser one. The lightest level that allows the surgery to be done properly is the one to aim for.

These are answerable in a sentence each by anyone who works this way routinely, and the answers should be about your case rather than about the office.

• Will exhaled carbon dioxide be monitored continuously for my case, and with a sampling line suited to an open mouth?
• Who is monitoring me, what training do they hold, and is monitoring their only responsibility for the whole case?
• Which permit does this practice hold, and does the level you are recommending for me sit inside it?
• When was the last documented emergency drill, and who leads if something goes wrong during my case?
• Is there anything in my history, my airway or my medication list that makes the monitoring plan different from the routine one?
• Which anaesthetic agents will be used, and does the emergency drug set cover them?
• What criteria have to be met before I am discharged, and who confirms them?

A patient reading a monitor display for the first time is usually looking at a large number they assume is the important one. It is generally the oxygen saturation, and it is generally the slowest reading on the screen. The faster, less glamorous trace beside it is the one that would have told the room first, and the person assigned to watch it is the reason that matters.