Continuous SpO2 Monitoring in Wearables: Technical Challenges and Clinical Utility

Continuous SpO2 monitoring sounds straightforward: keep the red and NIR LEDs running, and report blood oxygen every second. In practice, it's one of the harder challenges in wearable biosensing — requiring careful engineering trade-offs between power consumption, accuracy, motion tolerance, and form factor. This article explains why continuous SpO2 is hard and where it genuinely adds clinical value.

Why Continuous SpO2 Is Harder Than Spot Measurements

Power Consumption

Running red and near-infrared LEDs continuously at sufficient drive current for adequate SNR consumes 10-50 mW — a substantial fraction of a smartwatch's total power budget. The typical Apple Watch battery (308 mAh at 3.7V = 1.14 Wh) would drain in 23-114 hours under continuous SpO2 LED operation alone.

Consumer devices solve this by:

• Measuring SpO2 only during sleep (when power is less constrained and motion is minimal)
• Using on-demand spot checks during waking hours
• Adaptive sampling: increasing frequency when anomalies are detected
• Duty cycling LEDs at lower average power while maintaining adequate sampling rates

True 24/7 continuous SpO2 monitoring requires either larger batteries (medical patches with 5-7 day life), more efficient LED drive circuits, or clever power management that trades time resolution for battery longevity.

Signal Quality During Motion

SpO2 accuracy during motion is significantly worse than at rest. Red and NIR light penetrate deeper tissue than green (used for heart rate), meaning they are affected differently by motion-induced tissue compression. Motion creates artifacts simultaneously in both channels, corrupting the R-value ratio calculation that produces SpO2.

Clinical pulse oximeters use sophisticated algorithms (Masimo SET, Nellcor OxiMax) to detect and reject motion-corrupted data. Consumer wearables have less mature motion-rejection for dual-channel (red+NIR) signals than for single-channel (green) heart rate signals — green channel motion rejection has had more development time and algorithmic investment.

The practical result: consumer wearables typically suspend SpO2 estimation during activity, reporting "SpO2 unavailable" during movement above walking pace. This is actually appropriate behavior — displaying a motion-corrupted value would be worse than displaying nothing.

Skin Tone and Perfusion Variability

As discussed in the wearable pulse oximeter guide, melanin absorbs red and NIR wavelengths, reducing pulsatile signal amplitude and potentially biasing R-value calculations. With continuous monitoring, this bias appears in every measurement — accumulating over thousands of SpO2 estimates overnight.

In low-perfusion states (cold environment, exercise peripheral vasoconstriction, post-surgical cardiovascular stress), pulsatile amplitude at the wrist or wrist-adjacent sites drops sharply. When the AC/DC ratio falls below roughly 0.01%, most PPG-based SpO2 algorithms lose lock and either output erroneous values or flag the reading as invalid.

When Is Continuous SpO2 Clinically Useful?

Overnight Sleep Apnea Screening

This is the most established clinical use case for continuous wearable SpO2. Sleep apnea causes cyclical oxygen desaturations, typically lasting 20-60 seconds and occurring multiple times per hour. These desaturations are the physiological hallmark that overnight oximetry uses for screening.

The oxygen desaturation index (ODI) — the number of SpO2 drops of ≥3% or ≥4% per hour — correlates strongly with the apnea-hypopnea index (AHI) from polysomnography. An ODI4 (4% desaturation threshold) ≥15 events/hour corresponds roughly to moderate-to-severe OSA (AHI ≥15).

Wearable continuous SpO2 monitoring during sleep (Oura Ring Gen 3, O2Ring, Wellue SleepU) can calculate ODI with reasonable accuracy. Accuracy is better with ring devices than wrist devices due to higher perfusion at the finger.

Limitation: overnight oximetry cannot distinguish between obstructive and central sleep apnea, and cannot quantify respiratory effort (needed for AASM scoring). It is a screening tool, not a diagnostic replacement for polysomnography.

High-Altitude Monitoring

At altitude above 3,000 meters, SpO2 can drop below 90% during sleep (due to periodic breathing) even in acclimatized individuals. Continuous SpO2 tracking allows trekkers and mountaineers to monitor acclimatization progress and identify individuals at risk for high-altitude pulmonary edema (HAPE) before symptoms appear.

Consumer wearables perform reasonably well for this application — accuracy at SpO2 90-97% is adequate, and the monitoring environment (sleep at base camp, minimal motion) plays to their strengths. Garmin's Pulse Ox feature and Apple Watch's nighttime SpO2 monitoring are used routinely by altitude athletes.

Important caveat: above ~4,500 meters, SpO2 values may drop into the 80-89% range, where consumer device accuracy is uncertain. For expedition mountaineering above 6,000 meters, clinical-grade sensors (Nonin GO2 handheld, or medical wristwatch oximeters) are more appropriate.

COPD and Home Monitoring

Patients with moderate-to-severe COPD benefit from monitoring SpO2 trends over days to detect exacerbations early. A sustained drop from personal baseline (e.g., from 95% to 91%) may precede symptomatic exacerbation by 24-48 hours, potentially enabling earlier intervention.

This application requires a validated medical device, not a consumer wearable. Several research programs and commercial services (Current Health, Biofourmis) use FDA-cleared continuous SpO2 patches for COPD remote monitoring. Insurance and reimbursement pathways for these devices exist under RPM (remote patient monitoring) CPT codes.

COVID-19 and Post-Viral Illness

The COVID-19 pandemic highlighted "silent hypoxemia" — SpO2 drops below 90% in patients who felt well enough to stay home. Home pulse oximetry monitoring became part of COVID management protocols in several countries.

For continuous monitoring of COVID-19 patients at home, a 510(k)-cleared continuous wrist or ring oximeter is appropriate. Consumer smartwatches are not validated for this application. Several studies during the pandemic showed consumer device SpO2 missed clinically significant desaturations that a bedside oximeter would have caught.

Engineering Solutions to Continuous SpO2 Challenges

Multi-Wavelength SpO2 Estimation

Using more than two wavelengths (red + NIR) for SpO2 estimation can improve accuracy and melanin robustness. Four or more wavelength combinations allow estimation of multiple hemoglobin species (HbO2, Hb, met-Hb, carboxyhemoglobin) and melanin concentration, with melanin serving as a correction factor.

Masimo's rainbow SET technology uses this multi-wavelength approach to measure 12 hemoglobin parameters including total hemoglobin (SpHb), carboxyhemoglobin (SpCO), and methemoglobin (SpMet) — going well beyond simple SpO2 in clinical utility.

Consumer devices are beginning to adopt 4-wavelength sensing. Apple's Series 9 uses green, red, infrared, and UV-A LEDs, though UV-A is primarily used for background calibration rather than SpO2 spectral decomposition.

Adaptive Sampling and Gap-Filling

Rather than true second-by-second SpO2, sophisticated wearables use adaptive sampling:

• High-frequency sampling (1 Hz) during sleep when pattern recognition triggers concern
• Low-frequency sampling (every 30 seconds) when stable patterns are detected
• Predictive gap-filling: interpolating SpO2 during brief motion artifacts using preceding trend and HR-derived respiratory rate

This approach preserves battery life while maintaining the clinically important ability to detect sustained desaturations.

AI-Enhanced Accuracy

Neural network-based SpO2 correction has moved from research to commercial implementation. Training on large multi-skin-tone datasets allows the algorithm to learn skin-tone-dependent correction factors without explicit melanin measurement. Apple reportedly uses this approach in Series 9 to reduce skin-tone bias in SpO2 estimates.

What "Continuous SpO2" Actually Means on Consumer Devices

When a device advertises "continuous SpO2 monitoring," check the fine print:

• Apple Watch: Background SpO2 checks occur periodically during sleep, not every second. On-demand checks available anytime.
• Garmin: "All-day SpO2 monitoring" means periodic measurements (every 5-15 minutes), with increased frequency during sleep. Not second-by-second.
• Oura Ring Gen 3: SpO2 measurement during sleep, reported as average and low values for the night.
• O2Ring / Wellue: True near-continuous monitoring (every second), with configurable alarm thresholds. Designed for overnight sleep monitoring specifically.

True second-by-second continuous SpO2 wearables in consumer form factors are rare because of the power constraints described above.

Internal Resources

See also: PPG SpO2 accuracy and limitations, wearable pulse oximeter guide, PPG sleep apnea detection, and clinical-grade vs consumer PPG wearables.

FAQ

Can smartwatches monitor SpO2 continuously? Most consumer smartwatches do not monitor SpO2 truly continuously (every second). They take periodic measurements during sleep (every few minutes) or on-demand spot checks. True continuous monitoring requires more power and is typically reserved for dedicated medical-grade devices or ring sensors during sleep.

Is wearable SpO2 accurate enough to detect sleep apnea? Ring-based continuous SpO2 devices can screen for sleep apnea via oxygen desaturation index (ODI) calculation with sensitivity of ~80-85% and specificity of ~75-80% for moderate-to-severe OSA. Wrist-based devices perform somewhat worse. Neither replaces polysomnography for OSA diagnosis, but they are useful for initial screening, especially to identify patients likely to benefit from formal sleep testing.

What is a dangerous SpO2 level on a wearable device? As a general guideline, sustained readings below 92% warrant medical evaluation. However, interpret low readings cautiously on consumer devices — verify with a validated clinical pulse oximeter if concerned. During sleep, individual desaturation events to 88-90% occurring frequently (ODI4 >15/hour) suggest significant sleep apnea.

How does altitude affect wearable SpO2 accuracy? Consumer wearable SpO2 is reasonably accurate at high altitude when saturation is above 90%. Below 90%, accuracy is uncertain for consumer devices. Most consumer device calibration curves are based on healthy sea-level populations, and hypoxic conditions can introduce additional systematic errors.

Why does my smartwatch stop showing SpO2 during exercise? Motion artifacts corrupt the dual-wavelength (red + NIR) SpO2 signal during vigorous exercise. Wearables detect motion via accelerometer and typically suspend SpO2 estimation when motion artifact would make readings unreliable. This is appropriate behavior — displaying a potentially erroneous SpO2 value during exercise could be misleading.

What continuous SpO2 device is best for COPD monitoring? For COPD management requiring clinical-grade accuracy, the Nonin WristOx2 Model 3150 (FDA 510(k)-cleared wearable wrist oximeter) and the O2Ring by Wellue (FDA-registered, clinically validated ring oximeter) are commonly used. For hospital-integrated continuous monitoring, Masimo Radius PPG provides the highest accuracy. Consult with your pulmonologist about which device is appropriate for your monitoring needs.