How a Continuous Oxygen Monitor Works and What to Know

A continuous oxygen monitor tracks the oxygen level in your blood or tissue in real time, usually by shining two wavelengths of light through your skin and analyzing how much light gets absorbed. The most common version, the pulse oximeter, exploits a straightforward physical fact: oxygen-rich hemoglobin and oxygen-poor hemoglobin absorb red and infrared light differently. That principle has made pulse oximetry one of the most widely used measurements in medicine, but the technology has real blind spots, and a growing family of alternative sensors is filling some of those gaps.

Two Wavelengths, One Ratio

Hemoglobin is the molecule in red blood cells that carries oxygen from your lungs to the rest of your body. When hemoglobin is loaded with oxygen, it absorbs more near-infrared light and lets more red light pass through. When it has released its oxygen, the pattern flips: it absorbs more red light and transmits more infrared.¹ A pulse oximeter contains a pair of tiny light-emitting diodes, one producing red light and one producing infrared. On the other side of the sensor sits a photodetector that measures how much of each wavelength makes it through (or bounces back from) your tissue.

The clever part is how the device isolates the signal from arterial blood. Arteries pulse with each heartbeat, so the amount of light absorbed rises and falls rhythmically. The monitor locks onto that pulsatile component and ignores the steady absorption from bone, skin, venous blood, and other tissue. By comparing the ratio of pulsatile red absorption to pulsatile infrared absorption, the device calculates your oxygen saturation, displayed as SpO₂. A reading of 95 to 100 percent is normal for most healthy people at sea level.

The idea dates back decades. In the early 1970s, Japanese engineer Takuo Aoyagi was working on a dye-dilution cardiac output device when he noticed that changes in oxygen saturation disrupted the pulse signal he was trying to cancel out. He realized the disruption itself could be used to compute saturation, and modern pulse oximetry was born.²

Transmittance Versus Reflectance Sensors

Traditional clip-on pulse oximeters work in transmittance mode: the light source sits on one side of a thin piece of tissue, and the detector sits on the other side. This is why the fingertip and earlobe are the go-to spots. The tissue has to be thin enough for light to pass all the way through.³

That limits where you can take a reading. You cannot transilluminate the forehead, the chest, or most of the torso. Reflectance oximeters solve this by placing the LEDs and the photodetector on the same side of the skin. Instead of measuring light that passes through, they measure light that bounces back after scattering through tissue and blood vessels. This opens up almost any body surface as a potential monitoring site and is the approach used in most wearable devices, from hospital forehead sensors to consumer smartwatches. Researchers have built flexible, printed reflectance sensor arrays capable of mapping oxygenation across a patch of skin, rather than just giving a single-point number.³

When Poor Blood Flow Undermines the Reading

Pulse oximetry depends on detecting that rhythmic pulse of arterial blood. Anything that weakens peripheral blood flow can shrink the pulsatile signal to the point where the device either gives an inaccurate number or fails to give one at all. Hypotension, hypothermia, vasoconstriction from medications, and severe shock all fall into this category.⁴ In intensive care, where patients frequently have compromised circulation, this is a well-recognized problem. The perfusion index, a number the oximeter itself can display, reflects how strong the pulsatile signal is and serves as a warning when readings may not be trustworthy.⁵

The good news is that newer pulse oximeters with more sophisticated signal-processing algorithms handle low-perfusion situations better than older models. A systematic review found that about three-quarters of the oximeters studied were still deemed accurate in patients with poor perfusion, and the modern devices outperformed older ones. The review also noted that earlobe placement tended to produce more reliable readings than the fingertip when circulation was marginal.⁶ Warming techniques applied to the monitoring site can also help restore signal quality in critically ill patients.⁵

The Skin Pigmentation Problem

One of the most discussed limitations in recent years is that pulse oximeters tend to overestimate oxygen saturation in people with darker skin. The issue is not subtle. Multiple studies have found that the rate of “occult hypoxemia,” where a patient’s true arterial saturation is dangerously low (below 88 percent) but the oximeter displays a reassuring number in the 92-to-96 percent range, occurs roughly two to three times more frequently in Black patients than in White patients.⁷

The root cause is that the calibration algorithms used in most commercial pulse oximeters were historically derived from study populations that were overwhelmingly light-skinned. Melanin in the skin absorbs light at certain wavelengths, and if that absorption is not properly accounted for, the ratio-of-ratios calculation gets skewed. This can have real clinical consequences: a patient may appear adequately oxygenated on the monitor while actually needing supplemental oxygen or escalated care. Regulatory agencies and device manufacturers have been under growing pressure to improve calibration across a range of skin tones, but at the time of writing, no commercially available device has fully eliminated the bias.

How Probe Placement Affects Speed and Accuracy

Where you clip the sensor matters, and not just for signal strength. Because the pulse signal has to travel from the heart to the monitoring site, there is a built-in delay. A sensor on the ear detects a drop in oxygen saturation before one on the hand, which in turn detects it before one on the foot. In one study of healthy volunteers exposed to brief hypoxia, the mean delay between the ear and the foot was about 63 seconds, while the ear beat the hand by only about 6 seconds.⁸ In a stable patient on supplemental oxygen, this lag is unimportant. In someone whose oxygen level is dropping quickly, a minute of delay could matter.

There is also a settling period when you first start monitoring. When patients were placed on supplemental oxygen, about 95 percent reached a stable SpO₂ reading within roughly three and a half minutes. When oxygen was removed, equilibration took a bit longer, around four and a half minutes.⁹ Clinicians know to wait before trusting a new reading after any change in oxygen delivery.

Transcutaneous Oxygen Monitoring, a Different Approach

Pulse oximetry tells you the percentage of hemoglobin carrying oxygen. Transcutaneous oxygen monitoring (TcPO₂) does something different: it measures the partial pressure of oxygen that has diffused through the skin, giving a localized picture of how much oxygen is actually reaching the tissue at the sensor site. The technology uses an electrochemical sensor, typically based on an amperometric (current-measuring) principle, placed directly on the skin.

For oxygen to diffuse through the skin’s outer layer in measurable quantities, the sensor needs to warm the tissue underneath it, usually to around 44 °C. This warming dilates local blood vessels and melts some of the lipids in the outer skin layer, both of which increase oxygen diffusion.¹⁰¹¹ The heating element is built into disposable sensor patches so the temperature stays consistent.

TcPO₂ is especially valuable in vascular medicine. If you are trying to decide whether a patient with severe peripheral artery disease has enough blood flow to heal a wound or keep a limb viable, SpO₂ from a fingertip tells you almost nothing about what is happening near the wound. TcPO₂ placed near the affected tissue gives a direct readout. In patients with chronic limb-threatening ischemia, both the absolute TcPO₂ value and changes from baseline at 24 hours after revascularization were significantly linked to wound healing and limb preservation, and the association appeared earlier and stronger than what ankle-brachial index measurements could provide.¹² Tracking TcPO₂ over time in those patients adds further predictive value: a rising trend in readings independently increased the odds of healing.¹³

Overnight Oximetry and Sleep Apnea Screening

One of the most common uses of continuous oxygen monitoring outside the ICU is overnight pulse oximetry to screen for obstructive sleep apnea. During apnea episodes, breathing stops briefly and oxygen saturation dips. The pattern of repeated dips throughout the night, counted as the oxygen desaturation index, closely mirrors the apnea-hypopnea index measured by a full polysomnography sleep study.¹⁴ Overnight oximetry is far cheaper and simpler than an in-lab sleep study, making it a useful triage tool. It is not a replacement for polysomnography when a definitive diagnosis is needed, but a high desaturation index can confidently flag patients who should be fast-tracked for further evaluation.¹⁵

Neonatal Oxygen Targeting

In neonatal intensive care, continuous oxygen monitoring carries uniquely high stakes. Premature infants are vulnerable to both too little and too much oxygen. Insufficient oxygen raises the risk of death and brain injury, while excess oxygen is linked to retinopathy of prematurity, a condition that can damage the developing blood vessels in the eyes. The margin between harmful hypoxia and harmful hyperoxia is narrow, and the only way to stay inside it is continuous SpO₂ monitoring with tight alarm limits.

Studies have shown that the amount of time a preterm infant spends outside the targeted saturation range is directly associated with worse outcomes, including threshold retinopathy and mortality.¹⁶ When NICUs refined their oxygen-targeting strategies over successive epochs, rates of death, retinopathy, and chronic lung disease all declined significantly.¹⁷ The continuous oxygen monitor is the backbone of that strategy; without it, clinicians would be flying blind in one of the most sensitive dosing decisions in medicine.

Alarm Fatigue in Hospitals

Continuous monitoring generates continuous data, and continuous data generates continuous alarms. In hospitals, the traditional approach is simple threshold alarms: if SpO₂ drops below a set number, the alarm fires. The trouble is that brief, clinically meaningless dips below the threshold happen constantly, especially in patients who are moving, coughing, or have a loosely placed sensor. The result is a flood of false alarms. Nurses and respiratory therapists hear so many alarms that they begin to tune them out, a phenomenon formally recognized as alarm fatigue and ranked as one of the most significant safety hazards in hospital environments.¹⁸

Smarter alarm algorithms are one proposed fix. Instead of triggering on a single threshold crossing, some systems look at trend data, rate of change, or sustained duration below a threshold before sounding a clinical alarm. The goal is to filter out the noise without missing a genuine desaturation event. This is still an active area of engineering, and different hospitals and device manufacturers have taken different approaches, but the underlying recognition that raw threshold alarms create more problems than they solve is now widespread.

Consumer Smartwatches and the Accuracy Question

Most major smartwatches now include an SpO₂ sensor, using the same reflectance principle described earlier but packaged into a wrist-worn form factor. They can take spot readings or, in some models, log your saturation throughout the night. Whether these devices are accurate enough to be clinically meaningful is a different question.

Validation studies following the basic framework recommended by the FDA and the International Organization for Standardization have found that some smartwatches perform respectably well during short-term hypoxemia testing.¹⁹ But “meeting FDA standards” for consumer wellness devices does not always mean rigorous clinical validation. In some cases it simply means the device is not harmful, without strong evidence of accuracy across all conditions.²⁰ Wrist-based reflectance oximetry is inherently noisier than a fingertip clip: the sensor can shift with wrist movement, ambient light can leak in, and the distance between the LEDs and the arterial blood is greater. For tracking general trends, like noticing a pattern of nighttime desaturations that warrants a doctor visit, a smartwatch can be a reasonable starting point. For making clinical decisions, medical-grade devices remain the standard.

Monitoring at High Altitude

Portable pulse oximeters have become a common item in the kits of mountaineers and trekkers. At altitude, the lower barometric pressure means less oxygen is available with each breath, and SpO₂ drops accordingly. Tracking that drop can provide insight into how well your body is acclimatizing. A review of studies on pulse oximetry at altitude found that SpO₂ consistently decreases with acute altitude exposure and partially recovers as acclimatization progresses. People who develop acute mountain sickness tend to show lower SpO₂ values than those who do not, and there is growing evidence that persistently low readings serve as a useful marker of inadequate acclimatization.²¹

The predictive value is strongest for moderate-to-severe mountain sickness: people who went on to develop more serious symptoms showed more severe and prolonged desaturation from the very beginning of altitude exposure compared with those who stayed well.²² For mild altitude sickness, though, the overlap in SpO₂ values between affected and unaffected individuals is too large for the number to be a reliable predictor on its own. A pulse oximeter at altitude is a useful addition to paying attention to your symptoms, not a substitute for it.

Microneedle Sensors and the Future of Tissue Oxygen Measurement

Both pulse oximetry and transcutaneous monitoring measure oxygen indirectly, one by looking at hemoglobin’s color and the other by catching oxygen that drifts through the skin. A newer class of sensors aims to measure oxygen more directly by physically entering the tissue. Researchers have developed microneedle-based electrochemical oxygen sensors, tiny platinum-tipped needles short enough to penetrate just the outer skin layers without reaching pain-sensing nerve endings. One recent design uses platinum nanoparticles on a microneedle array to detect dissolved oxygen in dermal interstitial fluid in real time, with a response time of about nine seconds and stable performance over 90 minutes of continuous wear.²³

A separate effort produced a slightly longer microneedle sensor designed for intramuscular tissue oxygen readings, capable of distinguishing between high-oxygen and low-oxygen states in living tissue, though there is no established gold standard for tissue oxygen pressure that would allow precise calibration in vivo.²⁴ These devices are still firmly in the research stage, but the trajectory is clear: wearable patches that continuously measure local tissue oxygenation could eventually complement or partially replace the indirect approaches used today, particularly for wound monitoring and sports physiology, where what you really want to know is how much oxygen is reaching a specific spot in your body.

How Oximetry Works in Animals

The same physics apply across species, but the engineering gets complicated. Hemoglobin in mammals absorbs light the same way regardless of whether it belongs to a human, a dog, or an impala, so the core principle holds. The practical challenge is anatomy. Pulse oximeters calibrated for human fingertips do not automatically work on a tongue, a tail, or an ear folded over a clip at a different thickness. An evaluation across dogs, cats, and horses found that accuracy and failure rates varied widely depending on both the oximeter model and the species.²⁵

Probe site matters even more in animals than in humans. In anesthetized dogs, probes placed on the tongue had far lower failure rates than those on the tail or tibia, and vasoconstriction from certain sedative drugs pushed failure rates up across all sites.²⁶ In wild impala immobilized for veterinary procedures, only the under-tail site produced accurate readings, and even that location lost reliability when saturation dropped below 90 percent.²⁷ Wildlife veterinarians working with chemical immobilization, where animals frequently experience low oxygen levels, treat pulse oximeter readings with appropriate caution and usually supplement with other monitoring methods.