Early post-anaesthesia recovery parameters – a prospective observational study

Abstract

Study objectives

To evaluate the patients’ baseline condition upon arrival in the PACU as a method of assessment of the quality of anaesthesia, and to establish a model for future comparisons.

Design

Prospective observational study.

Setting

Surgical PACU in an academic tertiary hospital.

Patients

All patients (n = 11,241) arriving in our hospital’s recovery units after elective surgery.

Interventions

In this observational study, clinical data, vital signs and comfort parameters were collected from surgical patients who arrived in the PACU. For each parameter, its frequency distribution or percentage of occurrence was determined.

Main results

The incidence of anaesthesia associated side effects such as hypoxemia, cold extremities, shivering and/or vomiting was 5%. The incidence of nausea, sore throat, headache and/or pruritus was 9%. Sore throat occurred in 4.8% of intubated patients, in 4% after laryngeal mask insertions and in 3.6% with no usage of any airway device. From all patients 48% had no pain at all (VAS = 0), 31% had low pain scores (VAS 1–3) and 16% had moderate (VAS 4–6) pain levels, while 5% indicated severe pain (VAS 7–10). 97.5% of patients were normothermic (35.0–37.3°C), 77% had normal heart rate (60–100 beats per minute) and 74% had normal systolic blood pressure (90–140 mmHg). After use of neuromuscular relaxants, moderate clinical signs of residual curarisation (2 of 3 clinical criteria positive) were observed in 1% and slight clinical signs of residual curarisation (1 of 3 clinical criteria positive) were observed in 22.4% of patients. These findings were collected in all patients, independently whether they had reversal of neuromuscular relaxation.

Conclusions

Compared with the figures published in the literature, we report a lower incidence and severity of anaesthesia related side effects, measured as baseline data, upon patients’ arrival in the post anaesthesia care unit. The baseline data may serve as a model to trigger specific interventions aimed at improving the quality of anaesthetic care, which could be assessed in future investigations

Introduction

Quality control (QC) in the field of perioperative medicine may be performed amongst other means, by assessment of post-anaesthesia recovery parameters [1, 2]. Our department started in 2012 a QC program focused on early postoperative recovery parameters, which can be employed as an evaluation of the quality of anaesthetic care [3]. We assessed and documented the postoperative recovery time frames and recovery parameters. In the post-anaesthesia care units (PACU) of our hospital, the QC assessment can be performed immediately after arrival of each patient from the operating theatre. Our QC program is not limited in time and covers all surgical patients except for those with orthopaedic surgery which are not performed in our hospital and those who bypass the PACU for various reasons. Recorded data allow for periodical analysis as well as specific assessments before and after quality related interventions. In this article we present the data collected from the first calendar year since the initiation of the QC program. The collected data will be used as a baseline model for comparisons with future data that will be collected after having performed interventions to improve anaesthesia service and quality.

Methods

We extracted the incoming data obtained from 1^st of January to 31^st of December 2012 from all patients who were subjected to surgery in this period and who had passed through a postoperative recovery unit. The assessments were performed by recovery unit nurses, who by their professional qualification are competent to perform these assessments as an integral part of their regular duty. All assessments were made during the first few minutes after arrival in the PACU and represent real time results from the patients’ status at the beginning of their postoperative care. The data were entered into the patient’s hospital records and are accessible for reading and processing by authorized persons. The assessed parameters can be generally classified as

• – Objective observational parameters (cyanosis, shivering, vomiting)

• – Objective measurable parameters (hypoxemia = SpO₂ < 90%, temperature, heart rate, blood pressure, Clinical Muscular Capacity Test [CMCT])

• – Subjective observational parameters (prevalence of nausea, sore throat, headache, pruritus)

• – Subjective measurable parameters (pain score, alertness/sedation score)

The type of anaesthesia was specified in the baseline data collected: either general anaesthesia (with tracheal intubation, laryngeal mask or face mask), or regional anaesthesia (epidural or spinal). To assess sedation level we used an observational score ranging from 1 to 4 (1 = awake, 2 = sedated but able to respond verbally, 3 = non-awake but showing defence reactions to painful stimuli, 4 = no reaction at all = coma). Patients with scores 1 or 2 could be further assessed for subjective parameters and were also able to perform a locally developed clinical muscular capacity test (CMCT). The CMCT consists of 3 tests to rule-out residual neuromuscular relaxation: 1. Patient’s ability to squeeze strongly the hand of the observer, 2. The ability to lift his head and hold it elevated for 5 seconds, and 3. The ability to swallow 20 ml of water without signs of dysphagia (choking or coughing). The ability to perform all three tests can be interpreted as a high likelihood for absence of postoperative residual curarization (PORC), the ability to perform two (of three) tests suggests a low probability of PORC and the ability to perform one out of the three test suggests a moderate probability of PORC [4–9].

For pain assessment the VAS (visual analogue scale) was employed with scores ranging from 0 = no pain to 10 = worst pain), while all other subjective parameters were noted as present or absent. The limits of normal values for all these parameters were set, according to international standards (Table 1).

Table 1.

Objective measurable parameters, their definitions for normal vs. outlying values and the sources from where they were extracted

Parameter (unit)	Lower limit	Upper limit	Source
Core body temperature (°C)	35.0	37.3	Shoemaker, Allen L. What’s Normal? Temperature, Gender, and Heart Rate. Journal of Statistics Education. 4, 2 (1996)
Heart rate (b.p.m.)	60	100	Arrhythmia. National Heart, Lung, and Blood Institute. http://www.nhlbi.nih.gov/health/dci/Diseases/arr/arr_all.html. Accessed Feb. 12, 2011
Systolic blood pressure (mmHg)	90	140	For hypotension: National Heart, Lung, and Blood Institute. 2013. http://www.nhlbi.nih.gov/health/dci/Diseases/hyp/hyp_whatis.html. Accessed Mar. 2013 For hypertension: New Hypertension Guidelines: JNC-7. Clinical Cardiology Alert July 2003

Data were collected in Microsoft Excel (Microsoft Office 2010, Microsoft Corporation, Redmond, WA, USA). The obtained parameters are presented as descriptive statistics (in numbers, percentage or values as suited). Comparisons have been made for suitable parameters among different anaesthesia techniques and airway securing methods by applying IBM® SPSS® Statistics (version 21, SPSS® Inc. Chicago, Illinois, USA) using ANOVA with Bonferroni correction due to multiple comparisons. P values = 0.05 were considered significant. The manuscript was prepared in accordance with the STROBE criteria [10].

Results

During the data acquisition period of the calendar year 2012, a total number of 11,241 postoperative entries into the recovery units were registered. This represents the majority of surgical cases that were encountered during the observation time. The data completeness was dependent on the actual workload of the nursing staff that had to assess the patients, upon their arrival in the recovery unit. Occasionally, drop outs occurred at times of high work load with multiple patients arriving at the same time. We found 41 assessments that were not complete, and therefore certain details from these cases are missing. 99.6% of the collected data were complete.

The vast majority were general anaesthesia cases (97.6%). Table 2 presents the details of anaesthesia techniques employed. Table 3 presents the objective observational parameters such as presence or absence of cyanosis, shivering and vomiting.

Table 2.

Distribution of anaesthesia techniques

Anaesthesia techniques	Loco-region alanaesthesia		General anaesthesia
n	306		10,935
percent (globally)	2.7		97.6
	Spinal	Epidural	Intubation	Laryngeal mask	Face mask	Analgo-sedation
n	279	27	9027	1715	64	129
percent (for all techniques)	2.5	0.2	80.6	15.3	0.6	1.2
percent (separately for loco-regional or general anaesthesia)	91.2	8.8	82.6	15.7	0.6	1.2

Table 3.

Objective observational parameters such as presence or absence of hypoxaemia, shivering and vomiting

Objective observational parameters	Cyanosis	Shivering	Vomiting	None of these problems
n	25	341	59	10,609
percent	0.2	3.0	0.5	94.7

The incidence of nausea, sore throat headache and pruritus is summarized in table 4. The distribution of sore throat as a function of the means of securing the airway revealed the following results: sore throat was found in 430 of 8896 intubated patients (4.8%), in 69 of 1709 (4%) of laryngeal mask uses, in 2 of 17 (12%) in combined uses of tracheal intubation plus laryngeal masks, and finally in 20 of 551 (3.6%) patients who had no airway device inserted at all. Remarkably, there was no statistically significant difference in the incidence of postoperative sore throat when comparing patients with different airway management or different anaesthesia techniques (i.e. general vs. regional).

Table 4.

Subjective observational parameters such as presence or absence of nausea, sore throat, headache and pruritus

Subjective observational parameters	Nausea	Sore throat	Headache	Pruritus	None of these problems
n	279	511	130	72	10202
percent	2.5	4.6	1.2	0.6	91.1

The postoperative sedation scores were distributed as follows: sedation grade 1 in 6705 cases (60%), grade 2 in 4536 patients (40%), and there were no cases of sedation grades 3 or 4. Therefore, all patients were able to describe subjective parameters (Figure 1). 5463 (48%) patients indicated no pain at all (VAS = 0. Another 3362 (31%) patients had a VAS of 1 to 3 which is considered acceptable [11]. The remaining 2416 (16%) cases had pain scores from 4 to 6 and 600 (5%) had scores from 7 to 10 (Figure 2). Pain scores higher than 3 were treated with analgesics accordingly.

Fig. 1.

Boxplot representation of sedation and pain levels in all patients displayed as median, first & third quartiles (box edges), 5^th and 95^th percentiles (full circles) and range (whiskers); units are cm of a verbal rating scale (VRS) for sedation, or cm of a visual assessment scale (VAS) for pain

Fig. 2.

Frequency distribution of immediate postoperative pain levels with distinction of severity categories

Core temperature was 36.3 ± 0.5 °C ( 97.5% patients had normal temperatures), heart rate was 75 ± 16 beats/min; systolic blood pressure 128 ± 20 mmHg; and diastolic blood pressure was 71 ± 13 mmHg (Figure 3). 37 (0.3%) patients had a temperature < 35°C and 242 (2.2%) patients had > 37.3°C. 1767 (16%) patients had a heart rate < 60/min, 1851 (16%) a heart rate > 90/min, (68% of patients had normal heart rates). There were 105 (1%) patients with a systolic blood pressure < 90 mmHg, 2766 (25%) patients with a systolic blood pressure > 140 mmHg (74% had normal values).

Fig. 3.

Boxplot representation of core temperature and hemodynamic values in all patients displayed as median, first & third quartiles (box edges), 5^th and 95^th percentiles (full circles) and range (whiskers); units for temperature is °C, heart rate is beats per minutes (b.p.m.) and systolic blood pressure is mmHg

The Clinical Muscular Capacity Test (CMCT) estimates the probability of a postoperative residual curarization (PORC) [4, 7]. Only a complete failure to pass the 3 tests raised the suspicion of PORC. 116 patients (1%) failed to perform 2 of the 3 subtests (which means moderate probability for PORC), and 2519 patients (22.4%) failed to perform 1 out of the three tests (which means low probability for PORC).

Discussion

We assume that the initial postoperative condition of a patient is highly dependent on the quality of the anaesthesia that he/she has previously undergone. However, we are aware of the fact that the early postoperative status of a patient is multifactorial and not every deviation of the “normal” can be attributed to the quality of the previous anaesthesia alone. Certainly, the type of intervention, the length and degree of the surgical stress as well as the patient’s general health condition interferes with his appearance in the recovery unit. Thus, we assume that when viewing a high number of cases, the frequency and impact of the mentioned co-factors might remain constant and therefore the anaesthesia related effects may become perceptible.

The urge for better quality in anaesthesia is an ongoing effort in all environments. Examples of substandard anaesthesia are: inadequate efforts to maintain a sufficient oxygenation, unstable hemodynamics that may be caused by inadequate volume replacement, insufficient postoperative analgesia, not providing PONV prophylaxis when indicated, inappropriate dosing of anaesthetics, residual sedation or PORC, lack of efforts to prevent hypothermia, extubation of a patient’s trachea without fulfilling extubation criteria, etc.

The underlying hypothesis is that uncontrolled or longer lasting substandard anaesthesia care may leave traces in the immediate recovery period and that therefore, an early assessment of the listed parameters eventually would become recognizable in the PACU. By measuring the prevalence of early postoperative deviations from normal, we might possess a tool to quantify the impact of corrective interventions as soon as they have been implemented.

We emphasize the importance of early postoperative QC assessments, since certain signs of sub-standard anaesthesia would rapidly vanish in the course of the postoperative period and eventually remain undetected. Even if these not-normal findings might diminish due to self-correcting mechanisms or after therapeutic interventions, they may harm the patient and influence mid- and long-term outcomes. Besides, an immediate improvement of the patient’s status is not guaranteed, and patients with more severe concomitant diseases might suffer more serious complications. Of course, if pathological findings normalize by time, their relevance is reduced, but on an individual level it can be expected that they might also be precursors of postoperative complications.

Another important benefit of a standardized early postoperative QC assessment protocol is the easier recognition of the patient’s initial status. This facilitates the ability of the nursing staff to define the starting point of the recovery surveillance and treatment and the next immediate necessary steps how to proceed further during the upcoming recovery period. In our case, this starting point was set at the first 15 minutes after arrival in the PACU. During this period, the nursing staff takes over the patient and is briefed by the anaesthesiologist regarding the details of the patient and the course of surgery and anaesthesia. During this period, the relevant parameters are observed and measured and the questionnaire has to be filled accordingly. If necessary also certain therapeutic interventions may be initiated without delay.

The overall count of anaesthesia techniques reveals that in our department, regional anaesthesia occurred only in 2.7%. Therefore the typical problems associated but not exclusively attributed to this technique such as headache (1.2%) and pruritus (0.6%) are also rarely found. The typical adverse effects associated with general anaesthesia such as sore throat (4.6%), nausea (2.5%) and vomiting (0.5%) add up to 7.6%, which in turn is still a rather low proportion. Incidence of postoperative sore throat is reported in the literature to be up to 40% [12–14], while nausea and vomiting (PONV) has occurred at a rate of 30% [15–17]. Viewing this, we can register our department as having a modest amount of PONV, which in this respect means that our anaesthetic management is fairly adequate in this respect. Another quite astonishing fact was that adverse events that are typical for a certain anaesthesia technique (e.g. sore throat for general anaesthesia with tracheal intubation, or headache after neuroaxial blockades) did not show a statistically significant difference in their frequency when general vs. regional anaesthesia was compared. This might be because the proportion of loco-regional anaesthesia was not sufficiently large to attain a perceptible size. However, the incidence of sore throat after general anaesthesia either with tracheal intubation or laryngeal mask (4.6%) was not different from loco-regional anaesthesia (4.7%), thus indicating that this low total incidence is independent of manipulations at the airway in our setting.

Concerning the outcome of objective and subjective observational parameters, we have a large proportion of patients with no problems in these two categories (95 and 91 percent respectively). These results are markedly better than in comparable surveys in the past where the immediate recovery was uneventful in 23% or 26% respectively [18, 19]. We also can state that by having no cases with grades 3 or 4 sedation levels, all patients showed adequate arousal when arriving at the recovery unit, which is better than reported elsewhere [20–22]. This is also true for the (herein unaccounted) cases of residual hypnotic effects or central cholinergic syndrome, which were treated and solved in the operation room and had not yet been delivered to the recovery units with an inadequate level of consciousness. Another important aspect concerning the prevalence of postoperative somnolence and sedation is the increasing proportion of patients with obstructive sleep apnea syndrome (OSAS) [23, 24], which makes early and conscious arousal even more important.

A less favourable picture resulted in the case of postoperative analgesia, which turned out to be insufficient (> 3 VAS) in 2416 patients thus representing 21% of all cases (Figure 2). This includes pain levels of moderate and severe intensity, a frequency which is higher than found in the literature [25–27]. Although in the total population the mean VAS score was only 1.8 (± 2.3), this good average should not obscure the fact that a considerable portion of patients arrived in the PACU while they were still suffering pain. Of course, these patients were subsequently treated immediately after their arrival, but one can assume that the intraoperative regimen concerning choice, dosage and timing of analgesic treatment was not tailored to prevail in a sufficient manner during the transition from the OR to the PACU. Here we see much room for improvement and certainly this aspect will represent a priority topic in future interventions.

The assessment of objective measurable parameters also revealed a mixed picture. Hypothermia was found in 0.3% of cases, which is less than reported in other centers [28–30]; however, comparability is reduced by largely different types of surgery. Interestingly hyperthermia occurred even more often (2.2%), which otherwise does not seem to have been noticed elsewhere (no reference found for this problem). Hyperthermia might occur due to not paying enough attention to the overdoing of warming, in particular in cases of longer surgery on a less exposed body surface, when heat can accumulate by time and override the initial loss of core temperature during anaesthesia.

The hemodynamic status upon arrival to the recovery room was normal in the majority of cases, having normal heart rates in 8699 (77%) and normal blood pressure in 8370 (74%) of patients. A larger part of the remainder was hyperkinetic and a smaller part hypokinetic. This is comparable to other reports on the problems caused by postoperative changes of hemodynamics [31–33]. The more common hyperkinetic states (tachycardia and hypertension) may be attributed to both concomitant circumstances: a) rebound effect of diminishing anaesthesia in patients with latent or insufficiently treated hypertension, and/or b) the occurrence of postoperative pain that was not adequately prevented by the previous intraoperative treatment [11]. The latter problem might be more pronounced in cases when the intraoperative analgesia was based predominantly on the ultra-short acting remifentanil and a timely inclusion of adjuvant analgesic medication prior or around emergence from anaesthesia was lacking [34, 35]. The issue of insufficient postoperative analgesia is multifactorial and the tailoring of an optimal analgesic treatment is a matter that surpasses the objectives of this preliminary assessment. The far lesser fraction of found hypokinetic patients (bradycardia 16%, systolic hypotension 1% and diastolic hypotension 4%) may be attributed either to residual effects of anaesthetics stronger than intended, or to inadequate volume replacement (or both) [36, 37].

In this investigation we focused on the early postoperative appearance of patients in order to judge the previous anaesthetic care. However, in a different setting of scope investigations, these results also can be used for the assessment of the recovery phase itself by using these data as a starting point for the onward period of surveillance, in particular for QC efforts in the frame of “Enhanced Recovery After Surgery” (ERAS) programs [38, 39]. An essential issue is hereby to find an adequate balance between the two conflicting interests: a) a higher data quality requiring a higher workload for the staff vs. b) an easier to perform brief and fast assessment of basic data with a lesser information content. Due to the fact that the workload in the recovery units is often high and rather unpredictable, we opted for the second principle. By this we have assumed that with a less extensive assessment we could include the maximum possible caseload and have a low amount of dropouts. Even with this rather limited effort per person, the average time to conclude a complete assessment of one patient was around 8 to 12 minutes.

Conclusions

We presented baseline results from a successfully initiated and practicable quality control survey that became part of our standard perioperative care. The questionnaire designed to include the largest possible population size had to be reduced in content and complexity to a reasonable minimum. This approach proved to be acceptable for the nursing staff as well as to our demands and expectations to assess anaesthesia quality parameters. Stratification of the patients according to their morbidity as well as the surgical interventions according to size and invasiveness will be evaluated for appropriateness and usefulness in order to extract more reliable information regarding the anaesthesia behind each case. Finally, the results from this first year of the QC program represent a solid baseline and viable starting point for future assessments and controlled interventions.