Warning signs under anaesthesia: reading the relationships between the data

A warning sign can be a number clearly outside its limits. It can also be a gradual change, a discrepancy between measurements, or a response that is out of keeping with the patient's context.

In my practice, I see difficulty relating heart rate, pulse, blood pressure and the pulse oximeter. Each variable is useful, but none tells the whole story on its own. Before you choose an intervention, decide what deserves attention and whether the information is credible. Then read the picture as a whole.

This article is for education, not a diagnostic or treatment protocol, so use it alongside clinical assessment and your practice's protocols.

A low heart rate after an alpha-2 agonist

After an alpha-2 agonist such as dexmedetomidine, a fall in heart rate may be an expected effect. To read it well, you need to know which drug was given, when and under what circumstances. Without that, it may cause concern or lead to decisions poorly matched to the case.

An expected effect can still cause harm, though. Studies in dogs show cardiovascular changes with dexmedetomidine, including reduced cardiac output. Preserved blood pressure alone does not guarantee adequate perfusion ^1,2.

So "is the heart rate low?" is only the first question. Look at the baseline, the progression, the rhythm, blood pressure, the pulse and the rest of the assessment.

I do not propose a safe heart-rate range after dexmedetomidine that applies to every patient. The same number can mean different things in different patients, with different drugs and at different stages of the procedure. Context helps you decide, but it justifies neither automatic treatment nor automatically setting concern aside.

When pulse, pressure and the oximeter disagree

Relate a change in the pulse to the other cardiovascular data. AAHA includes peripheral pulse assessment and capillary refill time as complementary elements when assessing inadequate blood flow ³. Palpation is still not a blood pressure measurement, and a single monitor value cannot replace clinical assessment.

Say blood pressure changes and the pulse feels different. Check the observations and look for an explanation that fits. If the data seem discordant, face the discrepancy and resist resolving it by picking the most reassuring value. The information may be incomplete, there may be a measurement problem, or we may be seeing different aspects of the cardiovascular situation. Keep all of these open.

The pulse oximeter's perfusion index, commonly displayed as PI, can add information. It is not a direct blood pressure measurement and, on its own, it is not proof of adequate systemic perfusion.

One study followed 22 healthy female dogs undergoing ovariohysterectomy. PI fell over the procedure without consistently tracking the other indicators assessed. The authors concluded that low PI in that setting might not represent peripheral hypoperfusion ⁴.

That finding breaks the easy logic of "the index fell, so blood pressure fell". The index still deserves attention, but its interpretation has limits. Relate the change to signal quality, the patient assessment, blood pressure and the pharmacological context. PI may prompt you to check, but on its own it does not point to a cause or a treatment.

Sometimes the pulse oximeter gives unstable readings or loses its signal. That is a problem of availability or reliability, and it deserves attention. We cannot assume the patient is stable because a reading has gone, or automatically conclude that hypotension or hypoxaemia is present. Assess the patient and the measurement together. Monitoring guidance stresses combining instrument data with clinical assessment ⁵.

A low SpO₂ and an unavailable SpO₂ are different situations, even when both need checking. A technical problem should not make us dismiss a clinical one, and clinical suspicion does not turn a poor signal into a valid measurement.

Gradual change and rising EtCO₂

I find gradual change is frequently undervalued. A small fall in blood pressure or rise in EtCO₂ may look minor at one moment but become meaningful when repeated.

To compare, you need times and credible measurements. Isolated values far apart in time can miss what happened between them.

A persistent trend calls for assessment in context, though not every movement needs correcting. Ask whether there was an expected change in the anaesthetic, the procedure or ventilation. Ask whether the patient shows other concordant signs.

The reverse also holds. If there is an acute, concerning change, do not wait for a trend. The present situation may need an immediate response, even with no earlier sequence to compare.

Rising EtCO₂ is a good example, because it can prompt a response aimed only at the number. Interpretation has to take in the patient, ventilation, the breathing circuit and the measurement. AAHA's hypoventilation guidance includes assessing anaesthetic depth and looking for causes of rebreathing, such as breathing-circuit or CO₂ absorbent problems, according to the system used ⁶.

If the person monitoring can see the capnogram, looking at it is part of the assessment. Analysing the waveform takes more than seeing the number. Choosing an explanation takes more than one variable: gather the information that shows why EtCO₂ changed, then assess the response to whatever is done.

More data, the same attention

The more relevant clinical information we have, the more we may have to go on. Adding numbers does not by itself guarantee control of every condition, or greater safety. You need to know what each variable measures, its limits and how it relates to the others. Wrong data, misaligned times and old values shown as current can make interpretation harder.

A missing variable should also be visible, because an empty field is not a silent confirmation that all is normal.

Reading these relationships takes time and free attention. In real teams, the person monitoring the anaesthetic may be answering the surgeon, filling in the chart or looking for information on another monitor.

Manual recording takes attention too: you find the variable, find the time and mark the right point on the graph. This comes from my own experience rather than a quantitative comparison between methods. Showing the sequence clearly may help with that work, but someone still has to assess the patient and act.

VetNoa is being developed to help capture data from existing monitors and organise it over the course of a case. It reads only the numbers shown on the monitor screen. It does not analyse waveforms: the capnogram, ECG, plethysmogram or spirometry loops. Waveform analysis is a later goal. Capture is periodic rather than continuous, so events can happen between images and go unrecorded. Capture quality and gaps have to be taken into account. The aim is to help the team look up and combine the information available. We do not claim it beats human attention or replaces clinical vigilance. It has no validated ability to predict deterioration.

Questions for reviewing a case

• Is the current value credible, and does it reflect the present situation?
• How has it changed, and what data are missing over that interval?
• Which drugs were given, and when?
• Are pulse, rhythm, blood pressure and clinical assessment consistent with each other?
• Is the change in PI being treated as complementary information?
• Are we telling a low SpO₂ apart from an unavailable signal?
• What do the waveforms the monitoring person can see show?
• Does the situation need immediate help or intervention?
• What was done, and how did the patient progress afterwards?