Study Design

Reduced Quality Factors

0 = no serious, - = serious,

- - = very serious

Positive Quality Factors

0 = none, + = one, ++ = multiple

Dichotomous Outcome Summary

Non-Dichotomous Outcome Summary

Brief description

Overall Quality

High, moderate, low,
very low, none

No of studies

Study Type

RoB

Indirectness

Imprecision

Inconsistency

Large Effect

Dose-Response

Confounder

# Intervention with Outcome

# Control with Outcome

RR (95% CI)

Outcome 1: Favorable neurologic outcome

1

OB

0

- -

-

0

0

0

0

Very low

2

EXP

0

- -

-

0

0

0

0

Very low

Outcome 2: Survival to discharge

1

CT

-

- -

0

0

0

0

0

Very low

5

OB

- -

-

0

0

0

0

0

Very low

Outcome 3: ROSC - none

Outcome 4: Identification of CPA

2

CT

0

- -

0

-

0

0

0

Very low

6

OB

0

- -

0

-

0

0

0

Very low

5

EXP

0

- -

0

0

0

0

0

Very low

Outcome 5: Time to start CPR – none

Introduction

Pulse oximetry is used to report arterial oxygen saturation of hemoglobin and thus estimate PaO2 in dogs and cats. Most pulse oximeters also provide an auditory signal indicating pulse detection. Because hypoxemia, resultant desaturation, and diminishing perfusion can precipitate CPA, pulse oximetry monitoring of at-risk animals may prompt caregivers to intervene prior to the occurrence of CPA. This PICO question was designed to determine whether monitoring dogs and cats at risk of CPA with pulse oximetry could improve outcome.

Consensus on science

Outcome 1: Favorable neurologic outcome

For the most critical outcome of FNO, we identified 1 observational study in people1 (very low quality of evidence, downgraded for very serious indirectness and for imprecision) and 2 experimental studies in dogs2,3 (very low quality of evidence, downgraded for very serious indirectness and imprecision) that addressed the PICO question.

Observational study – post-ROSC:

A single registry-based observational study of OHCA in 9,405 adults showed that in people who achieved ROSC in the field who were monitored with SpO2 in the pre-hospital PCA phase, SpO2 < 94% was associated with worse FNO at 30 days (RR 1.108, CI95 1.069,1.147] and SPO2 of 99-100% did not appear to be harmful (RR 0.9851, CI95 0.956-1.015) on univariate analysis. Multivariate analysis of 4,897 subjects confirmed that SpO2 < 94% portended a poorer FNO prognosis (aOR 1.39, CI95 1.02-1.89; P = 0.04) and that SpO2 > 98% was not a risk factor for poor FNO (aOR 0.85, CI95 0.70-1.03; P = 0.10).1

Experimental studies – both post-ROSC:

An experimental model of fibrillatory arrest in dogs showed that titrating oxygen supplementation following ROSC to an SpO2 of 94 – 96% for 1 hour resulted in improved FNO (P < 0.05) and fewer histopathologic brain lesions at 24 hours when compared to routine application of FiO2 1.0.3

A study with similar methodology evaluated experimental dogs’ brain histopathology at 24 hours post-ROSC and found more severe histopathologic lesions in the subjects’ brains that had not received oximetry-directed conditions for the 1 hour following ROSC.2 These experimental studies suggest that pulse oximetry monitoring may be useful in the post-arrest period to help avoid hyperoxemia, which may improve FNO.

Outcome 2: Survival to discharge

For the next most critical outcome of survival to discharge, we identified 1 clinical trial4 (very low quality of evidence, downgraded for serious risk of bias and very serious indirectness) and 5 observational studies5–9 (very low quality of evidence, downgraded for very serious risk of bias and serious indirectness) that addressed the PICO question.

Clinical trial:

One randomized, non-blinded clinical trial designed specifically to evaluate the utility of pulse oximetry for post-operative monitoring in 1219 adults in a post-surgical unit found no difference in survival to discharge whether people were monitored continuously with pulse oximetry following surgery or not.4

Observational studies in target species:

One retrospective case-control study in 237 dogs and 181 cats showed that absence of reporting SpO2 values during sedation or general anesthesia in cats was associated with death within 7 days; the same was not found for dogs.5 A retrospective case-control study in 635 dogs found no association between the use of a pulse oximeter for anesthetic monitoring and occurrence of death during or within 48 hours of the anesthetic procedure.6 A retrospective case-control study in 730 cats found that cats that had “pulse or pulse oximetry” monitoring during the procedure were three to four times less likely to die than those that did not (OR 0.2, CI95 0.1-0.4; P < 0.001).7 Oximetry monitoring was used in conjunction with other vital signs monitoring in all of these studies.

Observational studies in people:

A large retrospective cohort study in 96,512 people showed that differences in various vital signs including oxygen saturation (aOR 5.2, CI95 3.1-9.0; P < 0.001 for short-term death in patients with SpO2 < 90% on supplemental O2 compared to those with SpO2 > 95%) were independently associated with 1-day mortality. However, in this study, bradypnea (aOR 18.1, CI95 2.1-155.5; P < 0.008) was found to have stronger, and tachypnea (aOR 4.9, CI95 3.4, 7.3; P < 0.001) nearly as strong, an association with 1-day mortality than oxygen saturation.8 Similar findings were noted regarding 30-day mortality: aOR 3.7, CI95 2.8,5.0 (P < 0.001) for patients with SpO2 < 90% on supplemental O2 compared to those with SpO2 > 95%; and aOR 3.1 CI95 2.6, 3.6 (P < 0.001) for tachypnea.

Similarly, a smaller retrospective study of 358 people experiencing IHCA in a single hospital showed that the national early warning score (NEWS), which includes pulse oximetry values, could be used to predict survival at 30 days following CPA (medium NEWS OR 4.43, CI95 1.81,10.83; high NEWS OR 9.88 CI95 2.77,35.26 for non-survival at 30 days compared to a low NEWS).9 Oximetry monitoring was used in conjunction with other vital signs monitoring in both of these studies.

Outcome 3: ROSC – No studies identified

Outcome 4: Time to identification of CPA

For the next critical outcome of time to identification of CPA, we identified 2 clinical trials4,10 (very low quality of evidence, downgraded for very serious indirectness and inconsistency), 6 observational studies11–16 (very low quality of evidence, downgraded for very serious indirectness and inconsistency), and 5 experimental studies17–21 (very low quality of evidence, downgraded for very serious indirectness) that address the PICO question.

Clinical trials:

One clinical trial in people demonstrated that rescue events decreased from 3.4±2.2 (CI95 1.89,4.85) to 1.2±0.94 (CI95 0.53,1.88;P = 0.01) per 1,000 patient discharges after implementation of continuous pulse oximetry monitoring of postoperative patients in a general surgery ward.10

Another randomized, non-blinded clinical trial designed specifically to evaluate the utility of continuous pulse oximetry for post-operative monitoring in 1219 adults in a post-surgical unit showed no difference in transfer frequency from the unit into the ICU whether patients had SpO2 monitoring or not.4

Observational studies:

One observational study of hospitalized people found that 3816/13,115 (29.1%) patients experienced SpO2 < 80% in the 1 hour prior to IHCA.14

In a hospitalized pediatric population, oxygen desaturation was noted in 32% of 3647 medical emergency team events, with 6.1% (223 events) progressing to acute respiratory compromise and 0.5% (17 events) progressing to CPA. In this report, details of the cases that progressed to CPA were not specified.13

One observational study of 2179 people seen in an emergency department showed that of 551 people who were “up-triaged” due to worsening health status, 489 (88.7%) had an increased RR, and 539 (97.8%) had an increased RR or HR. Only 12 cases (2.2%) had normal RR and HR, who were up-triaged only due to abnormal SpO2. This observational, descriptive study concluded that RR and HR were more impactful than SpO2 in determining patient status.15

One retrospective study in 1980 hospitalized people found that SpO2 results from the emergency department did not distinguish which patients would require rapid response team activation in the first 72 hours of hospitalization (P = 0.076). Respiratory rate (aOR 1.92, CI95 1.38,2.67; P < 0.001) was the variable most strongly associated with rapid response team activation.11

Tangential but related observations in hospitalized people:

Another retrospective study in 833 postoperative human patients found that 37% experienced prolonged (≥ 1 hour) episodes of SpO2 < 90% and that 11% experienced at least 1 episode lasting ≥ 6 hours. Conversely, according to nursing records, clinical hypoxemia was estimated to have occurred in only 5% of patients per the q12 hour medical record.12

An observational study in 50 people following major abdominal surgery found that continuous SpO2 monitoring was superior for detection of desaturation events compared to intermittent evaluation. In this study, events of SpO2 < 92% with a duration of more than 60 minutes were observed in 58% of patients.16

Experimental:

In a human study of experimentally-induced upper airway obstruction, SpO2 was unreliable in reflecting upper airway airflow limitation; instead, respiratory rate, a visual analogue scale, and a dyspnea scale were statistically correlated with upper airway airflow limitation (p < 0.0001 for RR and p < 0.05 for visual analog and dyspnea scale scores).21

One experimental rabbit study that evaluated the SpO2 during progressive graded blood loss documented normal-appearing SpO2 tracings that were present until the mean arterial blood pressure had dropped to 44 mm Hg, indicating that SpO2 is not a reliable indicator of hypotension in this setting.17

An experimental swine study found that SpO2 monitoring failed to detect 3 minutes of complete upper airway obstruction in pigs pre-treated with 100% oxygen.19

However, an experimental study in dogs spontaneously breathing room air demonstrated that pulse oximetry is useful in this setting to detect tracheostomy tube obstruction greater than 25%, and the greater the degree of obstruction, the shorter the time required for SpO2 decline.18

An experimental study in swine showed increased bias as well as poor accuracy and precision of pulse oximetry during peri-arrest hypoxemia.20

Outcome 5: Time to start CPR – no studies identified

Treatment recommendation

In dogs and cats at risk of CPA (e.g., under anesthesia, in shock, in respiratory distress, post-ROSC), we recommend against monitoring only with a pulse oximeter.(strong recommendation, very low quality of evidence)

In dogs and cats at risk of CPA (e.g., under anesthesia, in shock, in respiratory distress, post-ROSC), we suggest continuous pulse oximetry monitoring in conjunction with continuous or frequent monitoring of other vital parameters such as respiratory rate, heart rate and rhythm, and arterial blood pressure.(weak recommendation, very low quality of evidence)

In cats under general anesthesia, we recommend continuous monitoring of pulse oximetry or pulse quality.(strong recommendation, very low quality of evidence)

In dogs and cats in which a pulse oximetry reading cannot be obtained and patient movement and non-patient factors are ruled out as the cause, we recommend assessment of perfusion status by other means (eg, pulse palpation, blood pressure measurement, ECG monitoring, apnea monitoring, plasma lactate concentration measurement, point-of-care cardiac ultrasound).(strong recommendation, expert opinion)

Justification of treatment recommendation

1. There is variable evidence in hospitalized people that pulse oximetry may contribute to alerting rescue teams to potentially critical situations. There was no clear evidence in veterinary species regarding pulse oximetry monitoring in pre- or post-operative patients with regards to the PICO question (other than anesthetized patients).
2. In anesthetized cats, record of pulse oximetry monitoring was associated with survival benefit compared to no such documentation. While this was not seen in a similar survey in dogs, no harm was associated with pulse oximeter usage in the canine population.
3. Continuous pulse oximetry can alert caregivers to the presence of brady or tachyarrhythmias, though ECG monitoring is likely superior for this indication (and probably less likely to be displaced than a pulse oximeter probe in veterinary patients).
4. Pulse oximetry is not a good indicator of hypotension, and so other monitoring should be used in patients where this is a likely to be of concern.
5. In patients receiving supplemental oxygen, meaningful respiratory events (eg, upper airway obstruction, apnea) may be missed with a pulse oximeter while PaO2 is still adequate.

Knowledge gaps

The utility of continuous or intermittent SpO2 measurement for identification of impending or occurring CPA in dogs and cats in a clinical setting is unknown.

The design of the currently available pulse oximeter probes do not make them amenable to continuous monitoring of awake veterinary patients; development of probes that can be used in a continuous manner in veterinary species is encouraged.