Abstract
Background. We have previously reported that stroke volume is reduced in a majority of elderly patients undergoing surgical repair of hip fracture before and after intrathecal injection of anaesthetic. We aimed to investigate these observations further in a prospective study of elderly patients undergoing elective hip or knee arthroplasty under spinal anaesthesia.
Methods. Patients ≥65 yr undergoing elective arthroplasty were monitored with LiDCOplus™ preoperatively (baseline), before and continuously for 45 min after spinal anaesthesia. Postspinal hypotension was defined as systolic blood pressure (bp) < 100 mm Hg or > 30% decrease from baseline. Associations between post-spinal hypotension and haemodynamic changes before (i.e. between baseline and before injection) spinal anaesthesia were analysed by logistic regression analysis.
Results. Twenty patients with a mean age of 74 (range 66–89) yr were included. Stroke volume index decreased by 14% (95% CI 9.3%–19%) before spinal anaesthesia. When patients were categorised according to post-spinal hypotension (Y/N) the patterns of haemodynamic changes differed. In the hypotensive patients, cardiac index progressively decreased whereas it increased initially in the non-hypotensive patients. Reduction of cardiac index from baseline before spinal anaesthesia was associated with increased risk of hypotension: OR 0.79 (95% CI 0.60, 0.91). The predictive value of reduced cardiac index was good (AUC under ROC curve 0.91).
Conclusions. A decrease in cardiac output from baseline before spinal anaesthesia and an inability to increase it after induction may be important features of postspinal hypotension in elderly patients.
Editor’s key points
Hypotension is common in patients undergoing spinal anaesthesia, especially in the elderly.
In this small preliminary observational study of elderly patients undergoing elective lower limb hemiarthroplasty, cardiac index and stroke volume index (SVI) decreased before anaesthesia.
These changes were more marked and persisted in patients who became hypotensive after spinal anaesthesia.
In contrast, cardiac index and SVI were maintained or increased in patients who did not subsequently become hypotensive.
It may be possible to identify patients at risk of post-spinal hypotension but more data are needed before drawing firm conclusions.
Methods
Study design and setting
This prospective exploratory observational study was conducted at the Section for Orthopaedic Surgery, Department of Anesthesiology and Intensive Care at Karolinska University Hospital Huddinge, Stockholm, Sweden.
The study complied with the Declaration of Helsinki and was approved by the Regional Ethics Review Board in Stockholm (ID 2010-2042-31/1). Written, informed consent was obtained from all participants.
Participants
We included patients aged 65 yr or older undergoing hip or knee arthroplasty in spinal anaesthesia. Exclusion criteria were: (i) weight <40 kg; (ii) concomitant lithium medication; (iii) informed consent not obtained; (iv) research team unavailable. Full details on patient selection are provided in the Supplementary Appendix (Supplementary Fig. S1).
Outcomes
Outcome measures were absolute and relative changes from baseline of haemodynamic variables before and after spinal anaesthesia.
Anaesthetic management
Patients were premedicated with acetaminophen 1–2 g and modified release oxycodone 5–10 mg. Medication with angiotensin converting enzyme (ACE) inhibitors and angiotensin-II-receptor blockers were stopped on the day of surgery while calcium channel blockers and beta blockers were continued. After establishing i.v. access, pre-anaesthetic fluid loading of 0–500 ml of Ringer’s acetate was given at the discretion of the attending anaesthetist, followed by buffered glucose 25 mg ml−1 at 1 ml−1 kg−1 h−1 and Ringer’s acetate at 2 ml−1 kg−1 h−1 during spinal anaesthesia and surgery.
Spinal anaesthesia was induced with isobaric bupivacaine 5 mg ml−1 and morphine 0.4 mg ml−1 injected intrathecally through a 25- or 27-G cannula inserted at the lumbar level. Sensory block height was tested by cold discrimination every 15 min. Goals for perioperative management included mean bp 70–110 mm Hg, ≥10 g dL−1 haemoglobin concentration and SaO2>95%. Additional premedication, intrathecal injection volumes and haemodynamic support after spinal anaesthesia by fluid coloading or vasopressor administration were given at the discretion of the attending anaesthetist.
Haemodynamic measurements
An 18-G or larger cannula was inserted into the antecubital vein and the radial artery was cannulated with a 20-G catheter after subcutaneous injection of local anaesthesia (lidocaine 1%). The LiDCOplus™ monitor was calibrated a minimum of two times according to the manufacturer’s instructions with a bolus of 0.3 mmol lithium chloride (0.45 mmol in patients >90 kg) to determine cardiac output from the indicator dilution curve and to obtain a calibration factor used for continuous measurement of haemodynamic variables by the pulse power analysis integrated in the LiDCOplus™ system. Central venous pressure (CVP) was estimated to 7 mm Hg as given by the LIDCO™ software. Values for baseline and for T0, just before intrathecal injection, were defined by taking mean of monitor readings over 180 s at steady state. At baseline patients were in the supine position with 30° head of bed and at T0, patients had either sat up or were in lateral position, as decided by the anaesthetist. Continuous measurements were then carried out until 45 min after the induction. Timepoints T5–T45 represent means over 60 s at five-min intervals. The LiDCOplus™ monitor was concealed from the attending anaesthetic personnel at all times.
Hypotension (Y/N) was defined as systolic bp <100 mm Hg or >30% reduction from baseline at any time from intrathecal injection to end-of-data collection using the LiDCO monitor readings. The definition used corresponds to our earlier trial.5 Systolic, diastolic, mean and pulse pressures (PP) were also included in the analysis. Mean arterial pressure (MAP) is presented in tables and figures of the manuscript while the systolic, diastolic values can be found in Supplementary Table S2.
Statistical analysis
We presumed that 20 patients would be a convenient sample size for an explorative study. Haemodynamic data were extracted from LiDCO™ software (LiDCOviewPRO version 1.1, Cambridge, UK) and data sets were constructed in Excel for Windows (2007). Linear interpolation was performed for missing values. Data was analysed in Statistica (version 10 StatSoftInc, St Tulsa, OK, USA) and GraphPad Prism (version 5.00 for Windows, GraphPad Software, San Diego, California, USA). Logistic regression analyses were performed in SAS (version 9.3; SAS Institute Inc, Cary, NC, USA) and Statistica. Data was tested for normal distribution, and parametric or non-parametric tests were applied accordingly. Haemodynamic changes from baseline after pre-anaesthetic fluid loading before spinal anaesthesia were analysed with logistic regression. Haemodynamic changes over time after spinal anaesthesia were analysed with repeated-measures ANOVA (one- and two-way) with Dunnett’s multiple comparison test and Bonferroni correction. Univariate logistic regression analyses were used to assess the associations between hypotension (Y/N) and each of the haemodynamic variables (baseline values and the relative changes after pre-anaesthetic fluid loading, i.e. before spinal anaesthesia) and the volume of post-anaesthetic fluid loading after spinal anaesthesia. Univariate logistic regression analyses were also used to assess additional factors that could be associated with postspinal hypotension (age, body position at time of intrathecal injection, bupivacaine dose 15 mg or higher, propofol administration before spinal anaesthesia, sensory block height Th6 or over at 15 and 30 min after spinal anaesthesia, concomitant cardiovascular medication). If any one of these factors proved significant, interaction effects were tested together with the significant haemodynamic predictors in a multivariate model. The goodness of fit of the model was assessed by the Hosmer-Lemeshow test. Odds ratio (OR) estimates with profile-likelihood 95% confidence intervals were determined. Receiver operating characteristic (ROC) curves were constructed to assess the discriminative value of significant predictors. A value of P<0.05 was considered statistically significant. Continuous data are presented as median (range) and categorical data as numbers (%) or otherwise specified.
Results
Twenty-three patients were included in the study; three were later excluded from the analysis because of loss of haemodynamic data in one patient and late conversion to epidural and general anaesthesia in two patients, respectively. Full details on patient inclusion are provided in the patient flow chart in Supplementary Figure S1.
Patient characteristics and baseline haemodynamic data
Patient characteristics and baseline haemodynamic data are presented in Table 1.
Patient characteristics, baseline measurements and clinical data. Continuous variables are given as median (range) and categorical variables as numbers. P-POSSUM, Portsmouth Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity; HT, hypertension; DM, diabetes mellitus; CNS, disease of the central nervous system; ACEi, angiotensin converting enzyme inhibitors; A2RB, angiotensin type 2-receptor antagonist; BB, betablockers; Cab, Calcium channel blockers; Diur, diuretics; L, lumbar; Th, thoracic; CI, cardiac index; DO2I, oxygen delivery index; SVI, stroke volume index; SVRI, systemic vascular resistance index; DAP, diastolic arterial pressure; PP, pulse pressure; MAP, mean arterial pressure; HR, heart rate. No significant differences in the characteristics and parameters above were seen between patients with and without postspinal hypotension
| All n=20 | No postspinal hypotension n=10 | Postspinal hypotension n=10 | |
|---|---|---|---|
| Age yr (mean) | 74 (66–89) | 71 (66–79) | 77 (68–89) |
| Sex (F/M) | 14/6 | 5/5 | 9/1 |
| BMI kg/m2 | 26.5 (17–34) | 27 (23–34) | 24 (17–32) |
| ASA (I/II/III/IV) | (1/11/7/1) | 1/4/5/0 | 0/7/2/1 |
| P-POSSUM | 17.5 (13–31) | 16 (14–29) | 17 (13–31) |
| Organ dysfunction | |||
| HT/Lung/Heart/Renal/DM/CNS/Dementia | 16/4/7/1/4/2/2 | 8/2/3/1/1/1/0 | 8/2/4/0/3/1/2 |
| Cardiovascular medication | |||
| ACEi or A2RB/BB/Cab/Diur | 8/6/5/6 | 6/3/2/3 | 2/3/3/3 |
| Type of surgery | |||
| Hip/knee | 11/9 | 4/6 | 7/3 |
| Baseline haemodynamic measurements | |||
| DO2I ml min−1 m−2 | 455 (331–681) | 441 (331–681) | 472 (350–654) |
| Cardiac index ml min−1 m−2 | 2.6 (2.0–4.4) | 2.7 (2.0–4.4) | 2.5 (2.0–3.6) |
| SVI ml m−2 | 40 (29–60) | 48 (34–60) | 38 (29–52) |
| SVRI dyn s cm−5 m−2 | 2813 (1674–4343) | 2801 (1674–4334) | 2964 (1992–4343) |
| SAP mmHg | 160 (138–213) | 158(138–182) | 170 (147–212) |
| DAP mmHg | 64 (50–82) | 65 (59–82) | 63 (50–102) |
| PP mmHg | 95 (78–142) | 97 (87–115) | 103 (84–142) |
| MAP mmHg | 99 (85–144) | 92 (78–108) | 101 (85–144) |
| HR bpm | 68 (47–91) | 59 (47–81) | 75 (55–91) |
| Clinical data | |||
| Fluid preload (ml kg−1) (ml) | 2.3 (0–6.8) | 2.7 (0.85–6.85) | 0.85 (0–6.7) |
| Fluid coload (ml kg−1) (ml) | 3.8 (0.8–12.5) | 4.2 (0.85–8.6) | 3.75 (1.1–12.5) |
| Bupivacaine dose in SPA (mg) | 15 (12.5–20) | 15 (12.5–20) | 14 (14–17.5) |
| Level of injection (L2-3)/(L3-4)/(L4-L5) | 1/17/2 | 0/10/0 | 1/7/2 |
| Sensory block level 15 min after SPA | Th9 (L1–Th4) | 9 (L1-Th8) | 8 (Th10-Th4) |
| All n=20 | No postspinal hypotension n=10 | Postspinal hypotension n=10 | |
|---|---|---|---|
| Age yr (mean) | 74 (66–89) | 71 (66–79) | 77 (68–89) |
| Sex (F/M) | 14/6 | 5/5 | 9/1 |
| BMI kg/m2 | 26.5 (17–34) | 27 (23–34) | 24 (17–32) |
| ASA (I/II/III/IV) | (1/11/7/1) | 1/4/5/0 | 0/7/2/1 |
| P-POSSUM | 17.5 (13–31) | 16 (14–29) | 17 (13–31) |
| Organ dysfunction | |||
| HT/Lung/Heart/Renal/DM/CNS/Dementia | 16/4/7/1/4/2/2 | 8/2/3/1/1/1/0 | 8/2/4/0/3/1/2 |
| Cardiovascular medication | |||
| ACEi or A2RB/BB/Cab/Diur | 8/6/5/6 | 6/3/2/3 | 2/3/3/3 |
| Type of surgery | |||
| Hip/knee | 11/9 | 4/6 | 7/3 |
| Baseline haemodynamic measurements | |||
| DO2I ml min−1 m−2 | 455 (331–681) | 441 (331–681) | 472 (350–654) |
| Cardiac index ml min−1 m−2 | 2.6 (2.0–4.4) | 2.7 (2.0–4.4) | 2.5 (2.0–3.6) |
| SVI ml m−2 | 40 (29–60) | 48 (34–60) | 38 (29–52) |
| SVRI dyn s cm−5 m−2 | 2813 (1674–4343) | 2801 (1674–4334) | 2964 (1992–4343) |
| SAP mmHg | 160 (138–213) | 158(138–182) | 170 (147–212) |
| DAP mmHg | 64 (50–82) | 65 (59–82) | 63 (50–102) |
| PP mmHg | 95 (78–142) | 97 (87–115) | 103 (84–142) |
| MAP mmHg | 99 (85–144) | 92 (78–108) | 101 (85–144) |
| HR bpm | 68 (47–91) | 59 (47–81) | 75 (55–91) |
| Clinical data | |||
| Fluid preload (ml kg−1) (ml) | 2.3 (0–6.8) | 2.7 (0.85–6.85) | 0.85 (0–6.7) |
| Fluid coload (ml kg−1) (ml) | 3.8 (0.8–12.5) | 4.2 (0.85–8.6) | 3.75 (1.1–12.5) |
| Bupivacaine dose in SPA (mg) | 15 (12.5–20) | 15 (12.5–20) | 14 (14–17.5) |
| Level of injection (L2-3)/(L3-4)/(L4-L5) | 1/17/2 | 0/10/0 | 1/7/2 |
| Sensory block level 15 min after SPA | Th9 (L1–Th4) | 9 (L1-Th8) | 8 (Th10-Th4) |
Patient characteristics, baseline measurements and clinical data. Continuous variables are given as median (range) and categorical variables as numbers. P-POSSUM, Portsmouth Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity; HT, hypertension; DM, diabetes mellitus; CNS, disease of the central nervous system; ACEi, angiotensin converting enzyme inhibitors; A2RB, angiotensin type 2-receptor antagonist; BB, betablockers; Cab, Calcium channel blockers; Diur, diuretics; L, lumbar; Th, thoracic; CI, cardiac index; DO2I, oxygen delivery index; SVI, stroke volume index; SVRI, systemic vascular resistance index; DAP, diastolic arterial pressure; PP, pulse pressure; MAP, mean arterial pressure; HR, heart rate. No significant differences in the characteristics and parameters above were seen between patients with and without postspinal hypotension
| All n=20 | No postspinal hypotension n=10 | Postspinal hypotension n=10 | |
|---|---|---|---|
| Age yr (mean) | 74 (66–89) | 71 (66–79) | 77 (68–89) |
| Sex (F/M) | 14/6 | 5/5 | 9/1 |
| BMI kg/m2 | 26.5 (17–34) | 27 (23–34) | 24 (17–32) |
| ASA (I/II/III/IV) | (1/11/7/1) | 1/4/5/0 | 0/7/2/1 |
| P-POSSUM | 17.5 (13–31) | 16 (14–29) | 17 (13–31) |
| Organ dysfunction | |||
| HT/Lung/Heart/Renal/DM/CNS/Dementia | 16/4/7/1/4/2/2 | 8/2/3/1/1/1/0 | 8/2/4/0/3/1/2 |
| Cardiovascular medication | |||
| ACEi or A2RB/BB/Cab/Diur | 8/6/5/6 | 6/3/2/3 | 2/3/3/3 |
| Type of surgery | |||
| Hip/knee | 11/9 | 4/6 | 7/3 |
| Baseline haemodynamic measurements | |||
| DO2I ml min−1 m−2 | 455 (331–681) | 441 (331–681) | 472 (350–654) |
| Cardiac index ml min−1 m−2 | 2.6 (2.0–4.4) | 2.7 (2.0–4.4) | 2.5 (2.0–3.6) |
| SVI ml m−2 | 40 (29–60) | 48 (34–60) | 38 (29–52) |
| SVRI dyn s cm−5 m−2 | 2813 (1674–4343) | 2801 (1674–4334) | 2964 (1992–4343) |
| SAP mmHg | 160 (138–213) | 158(138–182) | 170 (147–212) |
| DAP mmHg | 64 (50–82) | 65 (59–82) | 63 (50–102) |
| PP mmHg | 95 (78–142) | 97 (87–115) | 103 (84–142) |
| MAP mmHg | 99 (85–144) | 92 (78–108) | 101 (85–144) |
| HR bpm | 68 (47–91) | 59 (47–81) | 75 (55–91) |
| Clinical data | |||
| Fluid preload (ml kg−1) (ml) | 2.3 (0–6.8) | 2.7 (0.85–6.85) | 0.85 (0–6.7) |
| Fluid coload (ml kg−1) (ml) | 3.8 (0.8–12.5) | 4.2 (0.85–8.6) | 3.75 (1.1–12.5) |
| Bupivacaine dose in SPA (mg) | 15 (12.5–20) | 15 (12.5–20) | 14 (14–17.5) |
| Level of injection (L2-3)/(L3-4)/(L4-L5) | 1/17/2 | 0/10/0 | 1/7/2 |
| Sensory block level 15 min after SPA | Th9 (L1–Th4) | 9 (L1-Th8) | 8 (Th10-Th4) |
| All n=20 | No postspinal hypotension n=10 | Postspinal hypotension n=10 | |
|---|---|---|---|
| Age yr (mean) | 74 (66–89) | 71 (66–79) | 77 (68–89) |
| Sex (F/M) | 14/6 | 5/5 | 9/1 |
| BMI kg/m2 | 26.5 (17–34) | 27 (23–34) | 24 (17–32) |
| ASA (I/II/III/IV) | (1/11/7/1) | 1/4/5/0 | 0/7/2/1 |
| P-POSSUM | 17.5 (13–31) | 16 (14–29) | 17 (13–31) |
| Organ dysfunction | |||
| HT/Lung/Heart/Renal/DM/CNS/Dementia | 16/4/7/1/4/2/2 | 8/2/3/1/1/1/0 | 8/2/4/0/3/1/2 |
| Cardiovascular medication | |||
| ACEi or A2RB/BB/Cab/Diur | 8/6/5/6 | 6/3/2/3 | 2/3/3/3 |
| Type of surgery | |||
| Hip/knee | 11/9 | 4/6 | 7/3 |
| Baseline haemodynamic measurements | |||
| DO2I ml min−1 m−2 | 455 (331–681) | 441 (331–681) | 472 (350–654) |
| Cardiac index ml min−1 m−2 | 2.6 (2.0–4.4) | 2.7 (2.0–4.4) | 2.5 (2.0–3.6) |
| SVI ml m−2 | 40 (29–60) | 48 (34–60) | 38 (29–52) |
| SVRI dyn s cm−5 m−2 | 2813 (1674–4343) | 2801 (1674–4334) | 2964 (1992–4343) |
| SAP mmHg | 160 (138–213) | 158(138–182) | 170 (147–212) |
| DAP mmHg | 64 (50–82) | 65 (59–82) | 63 (50–102) |
| PP mmHg | 95 (78–142) | 97 (87–115) | 103 (84–142) |
| MAP mmHg | 99 (85–144) | 92 (78–108) | 101 (85–144) |
| HR bpm | 68 (47–91) | 59 (47–81) | 75 (55–91) |
| Clinical data | |||
| Fluid preload (ml kg−1) (ml) | 2.3 (0–6.8) | 2.7 (0.85–6.85) | 0.85 (0–6.7) |
| Fluid coload (ml kg−1) (ml) | 3.8 (0.8–12.5) | 4.2 (0.85–8.6) | 3.75 (1.1–12.5) |
| Bupivacaine dose in SPA (mg) | 15 (12.5–20) | 15 (12.5–20) | 14 (14–17.5) |
| Level of injection (L2-3)/(L3-4)/(L4-L5) | 1/17/2 | 0/10/0 | 1/7/2 |
| Sensory block level 15 min after SPA | Th9 (L1–Th4) | 9 (L1-Th8) | 8 (Th10-Th4) |
Twenty patients, fourteen women and six men, predominately ASA II-III with a age of 74 (66–89) yr completed the study and were included in the analysis. When compared with the Swedish national registries for hip- and knee arthroplasty, the patients participating in this study corresponded well to the general population undergoing this type of surgery in terms of age, gender, BMI and ASA classification.7,8 A majority (N=16) had diagnosed hypertension and cardiac disease was present in a large proportion of participants (N=7). No patient had atrial fibrillation or other arrhythmias during the haemodynamic measurements.
Haemodynamic changes before intrathecal injection
Haemodynamic changes between baseline and intrathecal injection are presented in Table 2 for all patients together and for post hoc subgroups of patients with and without postpinal hypotension. Seventeen patients received pre-anaesthetic fluid loading. Significant changes occurred in all haemodynamic variables before intrathecal injection. SVI decreased in 18 patients by a mean of 5.8 ml (SD 4.1), the cardiac index decreased by 0.2 L min−1 (0.4) a mean and DO2I by 38 ml min−1 m−2 (72). Simultaneously, systemic vascular resistance index (SVRI) increased in 18 patients by 735 (492) dyn s cm-5 m−2, while heart rate (HR) increased in 16/20 patients by 5.1 (8.0) beats min−1. One patient was given 0.1 mg of i.v. phenylephrine three min before spinal anaesthesia; in the other patients no vasopressor was administered during this period. Body position just before intrathecal injection did not influence haemodynamic changes observed between baseline and time point T0 (Supplementary Table S1). Five patients in each group sat up and five were in the lateral position just before anaesthesia.
Haemodynamic changes (Δ) from baseline before SPA. Δ=T0–BL. T0 are means over 180 s just before injection of SPA and BL are means over 180 s at steady state at baseline after LiDCO calibration. Relative changes as expressed as mean change in % (95%, CI). CI, cardiac index; DO2I, oxygen delivery index; MAP, mean arterial pressure; HR, heart rate; SVI, stroke volume index; SVRI, systemic vascular resistance
| ALL Relative change, % (95%, CI) N=20 | No postspinal hypotension Relative change, % (95%, CI) N=10 | Postspinal hypotension Relative change, % (95%, CI) N=10 | Estimated OR for hypotension (95%, CI) | P-value | |
|---|---|---|---|---|---|
| DO2I | –7.7 (–15, –0.8) | 1.7 (−5.2, 8.6) | −17.2 (−26.6, −7.7) | 0.83 (0.67, 0.94) | 0.0182 |
| Cardiac index | –7.4 (–14, –0.5) | 2.4 (−3.9, 8.6) | −17.2 (−26.6, −7.7) | 0.79 (0.60, 0.91) | 0.0176 |
| SVI | –14 (–19, –9.3) | −9.3 (−16.2, −2.5) | −19.3 (−26.2, −12.4) | 0.89 (0.77, 1.01) | 0.0755 |
| SVRI | +27 (19, 35) | 24.0 (9.7, 38.2) | 29.7 (18.2, 40.5) | 1.0 (0.96, 1.08) | 0.471 |
| MAP | +14 (5.5, 23) | 21.5 (12.5, 30.6) | 6.6 (8.1, 21.3) | 0.94 (0.88, 1.01) | 0.104 |
| HR | +7.9 (3.1, 13) | 13.4 (8.5, 18.3) | 2.4 (−5.1, 9.8) | 0.84 (0.67, 0.97) | 0.0601 |
| ALL Relative change, % (95%, CI) N=20 | No postspinal hypotension Relative change, % (95%, CI) N=10 | Postspinal hypotension Relative change, % (95%, CI) N=10 | Estimated OR for hypotension (95%, CI) | P-value | |
|---|---|---|---|---|---|
| DO2I | –7.7 (–15, –0.8) | 1.7 (−5.2, 8.6) | −17.2 (−26.6, −7.7) | 0.83 (0.67, 0.94) | 0.0182 |
| Cardiac index | –7.4 (–14, –0.5) | 2.4 (−3.9, 8.6) | −17.2 (−26.6, −7.7) | 0.79 (0.60, 0.91) | 0.0176 |
| SVI | –14 (–19, –9.3) | −9.3 (−16.2, −2.5) | −19.3 (−26.2, −12.4) | 0.89 (0.77, 1.01) | 0.0755 |
| SVRI | +27 (19, 35) | 24.0 (9.7, 38.2) | 29.7 (18.2, 40.5) | 1.0 (0.96, 1.08) | 0.471 |
| MAP | +14 (5.5, 23) | 21.5 (12.5, 30.6) | 6.6 (8.1, 21.3) | 0.94 (0.88, 1.01) | 0.104 |
| HR | +7.9 (3.1, 13) | 13.4 (8.5, 18.3) | 2.4 (−5.1, 9.8) | 0.84 (0.67, 0.97) | 0.0601 |
Haemodynamic changes (Δ) from baseline before SPA. Δ=T0–BL. T0 are means over 180 s just before injection of SPA and BL are means over 180 s at steady state at baseline after LiDCO calibration. Relative changes as expressed as mean change in % (95%, CI). CI, cardiac index; DO2I, oxygen delivery index; MAP, mean arterial pressure; HR, heart rate; SVI, stroke volume index; SVRI, systemic vascular resistance
| ALL Relative change, % (95%, CI) N=20 | No postspinal hypotension Relative change, % (95%, CI) N=10 | Postspinal hypotension Relative change, % (95%, CI) N=10 | Estimated OR for hypotension (95%, CI) | P-value | |
|---|---|---|---|---|---|
| DO2I | –7.7 (–15, –0.8) | 1.7 (−5.2, 8.6) | −17.2 (−26.6, −7.7) | 0.83 (0.67, 0.94) | 0.0182 |
| Cardiac index | –7.4 (–14, –0.5) | 2.4 (−3.9, 8.6) | −17.2 (−26.6, −7.7) | 0.79 (0.60, 0.91) | 0.0176 |
| SVI | –14 (–19, –9.3) | −9.3 (−16.2, −2.5) | −19.3 (−26.2, −12.4) | 0.89 (0.77, 1.01) | 0.0755 |
| SVRI | +27 (19, 35) | 24.0 (9.7, 38.2) | 29.7 (18.2, 40.5) | 1.0 (0.96, 1.08) | 0.471 |
| MAP | +14 (5.5, 23) | 21.5 (12.5, 30.6) | 6.6 (8.1, 21.3) | 0.94 (0.88, 1.01) | 0.104 |
| HR | +7.9 (3.1, 13) | 13.4 (8.5, 18.3) | 2.4 (−5.1, 9.8) | 0.84 (0.67, 0.97) | 0.0601 |
| ALL Relative change, % (95%, CI) N=20 | No postspinal hypotension Relative change, % (95%, CI) N=10 | Postspinal hypotension Relative change, % (95%, CI) N=10 | Estimated OR for hypotension (95%, CI) | P-value | |
|---|---|---|---|---|---|
| DO2I | –7.7 (–15, –0.8) | 1.7 (−5.2, 8.6) | −17.2 (−26.6, −7.7) | 0.83 (0.67, 0.94) | 0.0182 |
| Cardiac index | –7.4 (–14, –0.5) | 2.4 (−3.9, 8.6) | −17.2 (−26.6, −7.7) | 0.79 (0.60, 0.91) | 0.0176 |
| SVI | –14 (–19, –9.3) | −9.3 (−16.2, −2.5) | −19.3 (−26.2, −12.4) | 0.89 (0.77, 1.01) | 0.0755 |
| SVRI | +27 (19, 35) | 24.0 (9.7, 38.2) | 29.7 (18.2, 40.5) | 1.0 (0.96, 1.08) | 0.471 |
| MAP | +14 (5.5, 23) | 21.5 (12.5, 30.6) | 6.6 (8.1, 21.3) | 0.94 (0.88, 1.01) | 0.104 |
| HR | +7.9 (3.1, 13) | 13.4 (8.5, 18.3) | 2.4 (−5.1, 9.8) | 0.84 (0.67, 0.97) | 0.0601 |
Haemodynamic changes after intrathecal injection
Details on clinical management are presented in Table 1. Individual haemodynamic changes with vasoactive administration overlaid and individual fluid pre-and coloading are presented in Supplementary Figure S2. Mean haemoglobin level decreased from 13.1 g dL−1 (sd 1.0) at baseline to 12.1 (1.2) g dL−1 at end-of-data collection before start of surgery (P<0.01). Hypotension developed in 10 patients (50%), according to the LiDCO monitor readings.
At the end of data collection, 45 min after spinal anaesthesia, cardiac index and DO2I had decreased by a mean of −19% (95% CI 10%, 27%) and SVI by −15% (9.6%, 21%) compared with baseline. Changes in DO2I over time after spinal anaesthesia followed the cardiac index changes in this analysis as baseline values for haemoglobin level and arterial oxygen saturation were used. SVRI decreased by −18% (95% CI 3.5%, 32%) ten min after spinal anaesthesia (T0–T10) but remained close to baseline values thereafter. MAP had decreased by −17% (95% CI 11%, 22%) and HR decreased by −4.8% (95% CI 0.3%, 9%) from baseline at the end of data collection.
Exploratory analyses of hypotension
The hypotensive and non-hypotensive patient characteristics are given in Table 1. Of the 10 patients who had hypotension based on LiDCO readings, seven received vasoactive therapy with phenylephrine 0.1 mg (n=5) or ephedrine 5 mg (n=2), and six of these received more than one bolus (2–6 boluses). One patient in the non-hypotensive group was administered vasopressor. Median time to vasopressor administration was 8.5 (4–22) min. In hypotensive patients, cardiac index and SVI decreased simultaneously and progressively together with MAP (Fig. 1). In contrast, in non-hypotensive patients, we observed an initial increase of cardiac index (T0 to T25) followed by a reduction (T25 to T45). SVI changes over time demonstrated statistically significant differences between hypotensive and non-hypotensive subgroups, the latter remaining closer to baseline values. There were no significant differences between hypotensive and non-hypotensive patients for changes in SVRI. Changes in HR differed between groups at all times, with a smaller initial increase and a larger decrease at the end of data collection in patients with postspinal hypotension.
(a–e) Haemodynamic changes after SPA (T0), per cent change from baseline (BL) in patients developing postspinal hypotension (N=10) and not developing postspinal hypotension (N=10). Data presented as median and interquartile range. P-values denote significant changes between groups: *P<0.05; **P<0.01; ***P<0.001; ns, non-significant.
The incidence of hypotension was not significantly associated with age, beta-blocker or other cardiovascular medication, body position during intrathecal injection, dose of bupivacaine (15 mg or higher), high sensory block (Th6 or above) at 15 min after spinal anaesthesia or propofol administration before spinal anaesthesia (Supplementary Table S3). However, sensory block higher than Th6 at 30 min after spinal anaesthesia was associated with hypotension (OR 0.028, 95% CI 0.001, 0.264). The administered volume of pre- or post-anaesthetic fluid loading (ml kg−1) had no significant association with incidence of hypotension (Supplementary Table S3).
Baseline haemodynamic values did not show any association with hypotension (Supplementary Table S3). Some significant associations were found between hypotension and the haemodynamic changes seen between baseline and intrathecal injection (Table 2). The changes in cardiac index and DO2I showed significant associations with hypotension while changes in HR and SVI did not. The sensory block level ≥Th6 at 30 min that was identified as a potentially associated factor showed no significant interaction effect in a multivariate model. The discrimination of relative changes in cardiac index and DO2I between baseline and T0 to predict postspinal hypotension was very good, illustrated by the areas under ROC curve (Fig. 2).
Receiver operating characteristics for the predictive value of percentage change of oxygen delivery index (DO2I) (A) and cardiac index (B) before spinal anaesthesia for the development of postspinal hypotension.
Discussion
In this cohort of elderly patients undergoing elective surgery under spinal anaesthesia, we found a decrease in SVI, cardiac index and DO2I before and after injection of spinal anaesthesia. This is consistent with the findings from our previous study in patients with hip fracture.6 Postspinal hypotension was associated with a progressive decrease of SVI and cardiac index, while there was no association with changes in SVRI. The reductions of cardiac index and DO2I observed before intrathecal injection were associated with higher risk of postspinal hypotension and showed good discrimination to predict hypotension.
Spinal anaesthesia has been associated with decreased cerebral blood flow and myocardial ischaemia after hypotension in the elderly.9,10 The association between mortality and hypotension was described not only in studies based on electronic data capture1 but also in studies based on audits (UK ASAP-2).2 In order to minimize hypotension after spinal anaesthesia in the elderly and frail it is important to characterize the haemodynamic changes and mechanisms.
In contrast to earlier studies, our data are prospective, indexed for body surface area, with baseline measurements temporally separated from start of anaesthesia and no exclusion of hypotension requiring intervention (i.e. describing haemodynamics in a common clinical situation).
Advances in minimally-invasive monitoring have led to further understanding of the haemodynamic effects induced by spinal anaesthesia in obstetric patients11 but the extraordinary physiology of pregnancy makes these results difficult to generalize. Previous studies on spinal anaesthesia in the elderly have monitored haemodynamic changes using non-invasive methods such as non-invasive bp4,12–16 and thoracic bioimpedance,17–21 while pulmonary artery catheter has been applied only in a small number of studies.22–24 Methods based on pulse contour or pulse power analysis offer precise time-resolution and are feasible in awake patients, yet studies using these methods to study the effects of spinal anaesthesia in the elderly are few.25–27
An initial increase in cardiac output after spinal anaesthesia is known to occur and has been described in previous studies of elderly patients.17,23,27 In our study, this increase was detected in patients who were haemodynamically stable after spinal anaesthesia, while patients who developed hypotension did not exhibit this biphasic change. We observed a statistically significant reduction of SVI before intrathecal injection, and a transient increase of SVRI and MAP by more than 10%. Patients who did not develop hypotension, increased their cardiac index before induction while the patients developing hypotension did not. It is possible that these changes could reflect an increase in left ventricular afterload induced by changes in body position and stress at the time of intrathecal injection. A decrease in venous return could also contribute to the observed decrease in cardiac index but unfortunately our study did not include measurements of filling pressures. However, the decrease in SVI and CI continued progressively after spinal anaesthesia in the hypotensive group while SVRI returned to values slightly below baseline. The decrease in SVRI during the first 10 min after spinal anaesthesia by approximately 20% is in line with earlier studies that demonstrate a decrease of 20–30% during this period after spinal anaesthesia.22,23,25,27 However, none of these studies present repeated temporally separated measurements before intrathecal injection.
No association was found between increasing age and post-spinal hypotension, possibly attributed to the small sample size. The observed association between SVI reduction (before intrathecal injection) and hypotension might be a sign of diastolic dysfunction. Diastolic dysfunction could presumably be present in many patients given the high age and frequency of hypertension.28 We suggest that diastolic dysfunction as a risk factor of postspinal hypotension should be addressed in future studies involving echocardiographic measurements. We did not find any difference in the amount of fluid received in patients developing postspinal hypotension. However, the range of administered fluids varied greatly between patients given the observational approach.
Autonomous nervous control has been hypothesized to be central to development of hypotension after spinal anaesthesia. HR variability expressed as low to high frequency ratio has been significantly associated with postspinal hypotension.29 Higher baseline HR has previously been shown to predict postspinal hypotension in obstetric patients.30 Meyoff and colleagues31 identified baseline bp variability low-frequency band power (BPV LF) and reduction of near-infrared spectroscopy (NIRS) as predictors of hypotension after spinal anaesthesia in elderly patients.
This study has several limitations. It is small and underpowered as demonstrated by the lack of association between known risk factors for postspinal hypotension (i.e. higher volume of intrathecal local anaesthetic, level of blockade above Th 5).2,12 It was not designed to detect differences between subgroups of patients with or without hypotension and our exploratory analyses should be considered as hypothesis-generating. Given the observational design, anaesthetic interventions were provided at the discretion of the anaesthetist (i.e. fluid pre- and co-loading, vasopressor administration and sedative medication were not protocolized). Changes in haemoglobin levels and arterial saturation were not updated continuously leading to covariation of CI and DO2I during the postspinal measurements, and as haemoglobin decreased in most patients at the end of data collection the decrease in DO2I may have been even more pronounced. Also, SVRI was calculated by the LiDCOplus™ system using assumptions of central venous pressure making the influence of factors such as venous return and right side filling pressures difficult to estimate.
Despite these limitations, the present study describes haemodynamic changes in an uncontrolled clinical setting and suggests mechanisms that may underlie postspinal hypotension in the elderly which need further studies.
Conclusions
Our previously reported observation of haemodynamic changes before spinal anaesthesia in elderly hip fracture patients is here reproduced in a cohort of elderly patients undergoing elective surgery. A decrease in cardiac output before spinal anaesthesia and an inability to increase it after injection may be important features of postspinal hypotension in elderly patients. Further studies are needed for appropriate haemodynamic optimisation strategies based on physiological mechanisms and assessment of preanaesthetic predictors for hypotension in this population.
Authors’ contributions
Study design/planning: E.B., S.H.K., J.J., M.L-L. Study conduct: J.J., M.L-L., J.J., E.B. Data analysis: J.J., E.B., S.H.K. Writing paper: J.J., E.B., S.H.K. Revising paper: all authors
Supplementary material
Supplementary material is available at British Journal of Anaesthesia online.
Acknowledgements
Magnus Backheden, statistician at the Department of Learning Informatics, Management and Ethics (LIME) at the Karolinska Institute, aided in designing and performing the logistic regression analyses.
Declaration of interest
None declared.
Funding
This work was supported by the Stockholm County Council, project number 20140181.
