Impact of spinal anaesthesia and obesity on maternal respiratory function during elective Caesarean section*


  • *

    Presented in part at the European Society of Anaesthesiologists' Annual Meeting, Glasgow; June, 2003.

Dr Britta von Ungern-Sternberg


Spinal anaesthesia for Caesarean section has gained widespread acceptance. We assessed the impact of spinal anaesthesia and body mass index (BMI) on spirometric performance. In this prospective study, we consecutively assessed 71 consenting parturients receiving spinal anaesthesia with hyperbaric bupivacaine and fentanyl for elective Caesarean section. We performed spirometry during the antepartum visit (baseline), immediately after spinal anaesthesia, 10–20 min, 1 h, 2 h after the operation, and after mobilisation (3 h). Baseline values were within normal ranges. There was a significant decrease in all spirometric parameters after effective spinal anaesthesia that persisted throughout the study period. The decrease in respiratory function was significantly greater in obese (BMI > 30 kg.m−2) than in normal-weight parturients (BMI < 25 kg.m−2), e.g. median (IQR) vital capacity directly after spinal anaesthesia; −24 (−16 to −31)% vs. −11 (−6 to −16)%, p < 0.001 and recovery was significantly slower. We conclude that both spinal anaesthesia and obesity significantly impair respiratory function in parturients.

For the majority of anaesthetists, spinal anaesthesia (SA) has become the preferred anaesthetic technique for elective Caesarean section. SA has a higher anaesthetic success rate than epidural anaesthesia [1] and reduces the potential for maternal morbidity and mortality related to airway complications associated with general anaesthesia [2]. Despite some degree of motor blockade, SA has only a slight effect on spirometric volumes in normal weight, non-pregnant patients because of diaphragmatic compensation acting as a counterbalancing mechanism; SA is associated with a slight decrease in vital capacity (VC) of about 10% in normal weight patients [3, 4]. However, this may differ in obese individuals, as SA tends to decrease lung volumes to a greater extent in these individuals than in normal weight, non-pregnant patients [5]. Furthermore, obesity has a significant impact on the respiratory function of non-pregnant patients undergoing breast surgery or lower abdominal laparotomy, as indicated by a mean decrease in VC of 40% (SD 19%) in obese (BMI > 30 kg.m−2) vs. 12% (SD 7%) in normal weight (BMI < 25 kg.m−2) patients following general anaesthesia [6]. In industrial countries, there has been an increase in the prevalence of obesity in the general population as well as in pregnant women [7]. As pregnancy itself is associated with many changes in respiration that also impinge on respiratory function [8], obese pregnant women are likely to be at an increased risk of impaired respiratory function. In particular, the cephalic shift of the diaphragm caused by the expanding uterus is enhanced in the presence of obesity and jeopardises pulmonary gas exchange by a marked reduction in functional residual capacity [9] and a simultaneous rise in closing volume that may exceed functional residual capacity in about 50% of pregnant women. This predisposes to airway closure in the normal tidal volume breathing range [10]. Therefore, we hypothesised that obese parturients would have a significantly greater decrease in spirometric volumes than normal weight parturients following SA and a slower recovery of respiratory function.


Following approval by the Ethics Committee of the University of Basel, Switzerland, and after obtaining written informed consent, we consecutively assessed 71 healthy (ASA physical status I–II) parturients with term pregnancies (36–42 weeks gestation) in this prospective study. Exclusion criteria were bronchial asthma requiring regular therapy, cardiac problems associated with dyspnoea or severe psychiatric disorder. Parturients had to be free of pain to be included in this study. This was defined as a score of ≤ 20 mm on a 100 mm visual analogue scale (VAS, where 0 mm represented no pain and 100 mm the worst possible pain).

For spirometric measurements, we used a Vitalograph 2120 (Vitalograph, Hamburg, Germany). We standardised the spirometric assessments with each parturient in a 30° head-up position and in the absence of pain (VAS≤20 mm). After a thorough demonstration of the correct usage of the device, we measured VC, forced vital capacity (FVC), forced expiratory volume in 1 s (FEV1), peak expiratory flow rate (PEFR) and mid-expiratory flow (MEF25−75) and calculated the FEV1/FVC ratio. We performed spirometry at least three times to meet the criteria of reproducibility as defined by the European Respiratory Society [11] and recorded the best measurement for further analysis. At the antepartum assessment, we measured the weight and height of each parturient to obtain their BMI. The antepartum spirometric assessment was used as baseline value (M0).

Patients received intravenous ranitidine 50 mg and metoclopramide 10 mg 60 min before the operation and 30 ml 0.3 m oral sodium citrate immediately before transfer to the operating theatre. We performed SA according to our routine using a 25-G pencil point needle. After identifying the subarachnoid space at the L3–L4 or L2–L3 interspace with the parturient lying in the right lateral position, we administered 0.5% hyperbaric bupivacaine 12.5 mg with fentanyl 10 μg. Thereafter, the parturient was turned on her back and left lateral uterine displacement was achieved by a wedge placed under the right hip and a 10° tilt of the operating table to the left. We assessed the level of sensory blockade using an ethyl chloride spray. As soon as the sensory level was above T5, we passively moved the parturient into a 30° head-up position in order to prevent cephalic spread of local anaesthetic and performed the second spirometric assessment (M1).

For postoperative pain relief, we gave intravenous increments of methadone 2 mg to achieve a VAS pain score of ≤ 20 mm while coughing. The total dose of methadone given to each patient was neither limited nor weight adjusted. Basic analgesia consisted of paracetamol 1000 mg rectally directly after the operation. We did not administer any local anaesthetic into the wound.

As soon as a VAS pain score of ≤ 20 mm was achieved, spirometry was performed for the third time (about 10–20 min after the operation, M2). Spirometric measurements were repeated 1 h (M3), 2 h (M4) and 3 h (after mobilisation, M5) after the operation and the cumulative methadone requirement was recorded at each assessment. Prior to the last measurement (M5), the parturients were mobilised out of bed and encouraged to walk a few steps in the recovery area (all about 5 min). We performed the last spirometric assessment (M5) again in a 30° head-up position to assess the influence of mobilisation on spirometric volumes. However, we did not measure spirometric volumes directly before and after mobilisation in order not to interfere disproportionately with the process of early bonding between the mother and her newborn.

To allow comparison of the parturients, the spirometric values were calculated as percentage change from the baseline value measured pre-operatively (M0). For statistical analysis, a repeated-measures analysis of variance (anova) was used. We used a Wilcoxon rank sum test to compare measurements between the BMI groups (BMI < 25, 25–30, > 30 kg.m−2). For post hoc comparisons, a Bonferroni test was used. The Spearman rank correlation test was used to assess the relationship between spirometric measurements and BMI. A p-value of < 0.05 was considered significant. For statistical calculations, we used Stat View for Windows (SAS Institute Inc., Cary, NC, Version 5.0.1).


Seventy-one parturients were included in this study: of these, six (9%) declined to continue and 65 successfully completed the study. Seven parturients (11%) were occasional to moderate smokers (< 10 cigarettes per day) and three (5%) were ex-smokers who stopped smoking before becoming pregnant. The majority of parturients (n = 55; 85%) had been non-smokers all their lives. Patient characteristics and obstetric details are summarised in Table 1.

Table 1.  Characteristics of parturients (n = 65). Values are median (interquartile range [range]) or number (%).
Age; year32 (25–39 [20–40])
Body mass index
 Before pregnancy24 (18–30 [17–41])
 Term pregnancy30 (23–38 [21–50])
Gestation; weeks39 (38–40 [36–42])
Primigravida/multigravida28 (43%)/37 (57%)

Surgical anaesthesia was achieved in all parturients with a median (IQR [range]) upper sensory level of T4 (T3–T5[T2–T5]) at M1. Postoperatively, the assessed median upper sensory levels for different times of spirometry were as follows: T5 (T3–T7[T2–T12]), T8 (T5–T11[T3–L5]) and T12 (T9–L3[T6–none]) 20 min (M2), 1 h (M3) and 2 h (M4), respectively, after Caesarean section. There was no residual sensory level 3 h after the operation (after mobilisation, M5). The mean duration of surgery was 53 (SD 10) min.

In all parturients, spirometric values were all within normal ranges. However, there was a significant decrease in all spirometric parameters after effective SA (Table 2). VC decreased significantly more in the obese (BMI > 30 kg.m−2) than in normal-weight parturients (BMI < 25 kg.m−2) (Table 3). This decrease persisted over the whole observation period. FVC and FEV1 changed in parallel with VC, but both PEFR and MEF25-75 showed a significantly greater reduction and slower recovery than the other parameters (Table 2). The FEV1/FVC ratio was not affected by SA and remained unchanged throughout the observation period. Three hours after the operation and after patient mobilisation, baseline conditions of spirometric values had not been re-established in any of the BMI groups of parturients (Tables 3 and 4). At each assessment following SA and Caesarean section, there was a significant negative correlation between BMI and the spirometric parameters (Table 5).

Table 2.  Results of spirometry in parturients receiving spinal anaesthesia for elective caesarean section. Absolute values and changes of vital capacity (VC), forced vital capacity (FVC), forced expiratory volumes in 1 s (FEV1, peak expiratory flow rate (PEFR) and mid-expiratory flow (MEF25−75). Values are median (IQR) or % decrease of preoperative value. All changes were statistically significant compared with baseline values (repeated measure anova, p < 0.001).
VC [I]FVC [I]FEV1 [I]PEFR [I.min−1]MEF25−75[I.s−1]
Pre-operative; M0 3.4 (2.7–4.1) 3.2 (2.5–3.9) 2.9 (2.3–3.6)389 (326–452) 4.1 (3.5–4.7)
After SA; M1 2.8 (2.2–3.4) 2.7 (2.0–3.4) 2.4 (1.8–2.9)266 (201–332) 2.9 (2.2–3.5)
% decrease from M018 (7–27)17 (7–27)18 (6–30) 30 (17–43)29 (16–43)
After surgery; M2 2.7 (2.0–3.4) 2.6 (1.9–3.3) 2.4 (1.7–3.0)276 (207–346) 2.9 (2.1–3.7)
% decrease from M018 (3–33)17 (2–33)18 (3–34) 28 (16–41)28 (15–41)
1 h; M3 2.8 (2.0–3.6) 2.6 (1.8–3.4) 2.3 (1.6–3.1)271 (200–341) 2.9 (2.2–3.6)
% decrease from M017 (0–34)18 (0–35)19 (0–38) 29 (16–41)29 (16–41)
2 h; M4 2.7 (2.0–3.5) 2.7 (1.9–3.4) 2.4 (1.7–3.0)278 (210–347) 2.9 (2.2–3.6)
% decrease from M017 (2–33)17 (2–31)18 (2–35) 27 (16–37)25 (14–36)
After mobilisation/3 h; M5 2.8 (2.1–3.6) 2.8 (2.0–3.5) 2.4 (1.7–3.1)281 (210–350) 3.0 (2.2–3.8)
% decrease from M014 (0–29)16 (1–31)17 (1–33) 26 (12–39)25 (10–40)
Table 3.  Results of vital capacity in parturients receiving spinal anaesthesia for elective caesarean section according to body mass index (BMI). Values are median (IQR) or % decrease of pre-operative value. All changes to baseline were significant (repeated measure anova), the significances of all values between BMI < 25 and >30 (Wilcoxon signed rank test) are indicated (* = significant, n.s. = not significant), p < 0.001.
BMI; kg.m−2Vital capacity
<25 (n = 9)25–30 (n = 22)>30 (n = 34)<25 vs. >30
  1. *SA, spinal anaesthesia.

Pre-operative M0 3.4 (3.0–3.7) 3.4 (2.8–4.0) 3.3 (2.5–4.1)n.s.
After SA*; M1 3.0 (2.9–3.1) 2.8 (2.3–3.3) 2.5 (1.9–3.1)*
% decrease from M011 (6–16)15 (11–20)24 (16–31)*
After surgery; M2 3.1 (2.7–3.4) 2.9 (2.5–3.3) 2.5 (1.9–3.0)*
% decrease from M0 8 (2–18)15 (8–21)27 (15–40)*
l h; M3 3.1 (2.9–3.4) 2.9 (2.4–3.4) 2.4 (1.8–3.0)*
% decrease from M0 9 (3–15)14 (8–20)28 (15–41)*
2 h; M4 3.2 (2.9–3.6) 3 (2.5–3.4) 2.4 (1.8–3.0)*
% decrease from M0 6 (3–10)14 (9–20)27 (16–39)*
After mobilisation (3 h); M5 3.3 (3.0–3.6) 3 (2.4–3.7) 2.5 (2.0–3.0)*
% decrease from M0 4 (2–9)12 (7–17)23 (11–35)*
Table 4.  Effect of mobilisation on vital capacity (VC), forced vital capacity (FVC), forced expiratory volume in 1 s (FEV1), mid-expiratory flow (MEF25−75) and peak expiratory flow rate (PEFR) on parturients with BMI < 25 (n = 9) vs. BMI > 30 (n = 34). Values are % decrease of pre-operative value, median (IQR). Statistical significance (p-value) within the groups as determined by repeated-measures analysis of variance.
 Body mass index
At 2 hAt 3 h/after mobilisationp-value At 2 hAt 3 h/after mobilisationp-value
VC6 (3–10)4 (2–9)0.005227 (16–39)23 (11–35)<0.001
FVC6 (2–10)4 (2–10)0.0226 (14–38)23 (12–36)<0.001
FEV18 (6–11)7 (1–12)n.s.29 (17–40)26 (13–39)<0.001
MEF25−7518 (10–25)15 (8–23)0.00932 (23–40)32 (22–42)0.0003
PEFR19 (10–28)16 (8–25)0.006634 (26–43)33 (23–44)<0.001
Table 5.  Correlation coefficients of body mass index (BMI) to vital capacity (VC), forced vital capacity (FVC), forced expiratory volume in 1 s (FEV1), mid-expiratory flow (MEF25-75) and peak expiratory flow rate (PEFR) at the different assessment times (after effective spinal anaesthesia ([SA], M1), 20 min (M2), 1 h (M3), 2 h (M4) after the operation and after mobilization (3 h, M5) as determined by the Spearman rank correlation test, p < 0.001.
Time of
After SA; M1−0.73−0.76−0.76−0.79−0.78
At 20 min; M2−0.74−0.75−0.74−0.84−0.85
At 1 h; M3−0.79−0.80−0.78−0.83−0.84
At 2 h; M4−0.80−0.81−0.80−0.83−0.83
After mobilisation (3 h); M5−0.77−0.77−0.80−0.81−0.82

No intravenous methadone was necessary for any parturient during or directly after the operation. The amount of methadone (median (IQR [range])), given intravenously was 0 (0–0 [0–4]) mg during the first hour after the operation, 6 (2–10 [2–10]) mg between the first and second hour, and 6 (3.5–8.5 [4–12]) mg between the second and third hour. This added up to a total dose of methadone of 12 (8–16 [8–20]) mg over the entire observation period. All parturients maintained an arterial oxygen saturation of ≥ 97%.


Obese parturients presenting for Caesarean section have a high risk of anaesthetic and obstetric complications that contribute to peri-operative morbidity and mortality. Although these parturients are particularly at risk for pulmonary complications resulting from changes induced by pregnancy and obesity, there are no reported controlled trials on the impairment of peri-operative respiratory function in this subgroup. There are some reports on the respiratory effects of SA and epidural anaesthesia for Caesarean section in normal weight parturients [12–16], and only a small study on the effect of SA in obese, albeit non-pregnant, patients [5]. Thus, the present prospective study was designed to evaluate both normal weight and obese parturients scheduled for elective Caesarean section.

In our study, baseline spirometric values were all within normal ranges for normal weight as well as obese parturients. Not unexpectedly, there were no signs of airway obstruction during the whole study period, as the FEV1/FVC ratio is not or only minimally influenced by obesity [6] and apparently by SA. Following institution of SA, the decrease in VC observed in parturients with a normal BMI was comparable to findings reported in other studies [3, 4, 12–14], whereas VC values were significantly lower in those with a BMI > 30 kg.m−2 (−11% vs. −24%) (Table 3). A significant negative correlation between spirometric parameters and the BMI persisted throughout the study (Table 5). Any calculation of BMI values in parturients is somewhat arbitrary, as the ‘overweight’ of a parturient is partly attributable to surplus body water content, the weight of the foetus plus the hydramnion, and not solely due to excess adipose tissue. This puts intrinsic limitations on any straightforward comparison with non-pregnant obese individuals. Nevertheless, we used the common BMI categories as we still consider this parameter to be the most reliable for evaluating the influence of body configuration, even those of parturients, on respiratory mechanics.

During quiet breathing, the diaphragm is the principal muscle of inspiration, whereas expiration is mainly passive. In contrast, forced expiration depends on the muscles of the abdominal wall and to a lesser extent on the intercostal muscles. SA with an upper sensory level of up to T4 for elective Caesarean section induces muscle paralysis of the abdominal and intercostal muscles. This muscle paralysis is associated with a reduction in abdominal resistance that allows, at least in normal-weight, non-pregnant women, the diaphragm to move more easily during inspiration to compensate for the loss in lung volumes attributable to SA [3]. However, this compensatory mechanism is not fully effective during pregnancy and is likely to be abolished by obesity. There is another explanation for the reduction in respiratory performance that we observed in obese parturients. Obesity predisposes to the formation of atelectasis per se and even more so after anaesthesia [17], further contributing to postoperative compromise of respiratory function. Hypothetically, the delayed recovery from the impaired respiratory function associated with SA and Caesarean section could be caused by recumbency, surgical manipulations during Caesarean section, and peri-operative volume shifts, all of which contribute to changes brought about by pregnancy, obesity and SA and ultimately result in a marked reduction of ventilated lung tissue.

Interestingly, there was no significant difference between the spirometric values obtained before and after delivery of the baby as assessed postoperatively, although this should have enabled the diaphragm to move more easily. The reduction in VC measured 3 h after Caesarean section following full recovery from SA was primarily related to the BMI as a predictive factor for postoperative atelectasis formation and not to other factors interfering with spontaneous respiration, such as postoperative pain.

This study emphasises the importance of early mobilisation as a powerful measure to restore lung volumes; respiratory function only improved after mobilisation. Although this improvement might be partially due to the natural time course, as the impairment following SA might eventually resolve by itself, we consider mobilisation the main cause of improvement of respiratory function. But to determine the exact cause of this improvement of lung volumes, a direct comparison between lung volumes measured immediately before and after mobilisation would be necessary. We observed no improvement of lung volumes following SA until the parturients were mobilised; there was no difference in VC values determined immediately after SA (M1) and 2 h after Caesarean section (M4), despite quite different sensory block levels (median T4 at M1 and T12 at M4). Although a high sensory block (T4) might be expected to be accompanied by a higher degree of lung volume impairment compared with a lower sensory level (T12), which theoretically should only minimally influence lung volumes (especially as motor blockade tends to be even lower than the sensory level), respiratory function did not correlate with the level of SA in the postoperative period. Improvement of lung volumes was only achieved after mobilisation of the parturients, especially in the obese, in spite of the continuous regression of the motor block during the observation period.

Out of bed mobilisation presumably resulted in reopening of some atelectasis, thereby recruiting lung tissue for effective gas exchange. All parturients had an arterial oxygen saturation of ≥ 97% during the entire observation period, while receiving oxygen 2 l.min−1 via nasal cannula. Three hours after the operation, there was a small but significant improvement in all spirometric parameters with the exception of FEV1 in normal weight parturients. This is in contrast to a previous study [12] in which MEF25−75 decreased by a further 10% from −18% after 2 h following Caesarean section to −28% after 4 h. This difference may be the result of variations in postoperative pain levels that increased from a mean (SD) of 33 (4.5) after 2 h to 47 (3.6) after 4 h, despite access to patient controlled analgesia (PCA) morphine [12]. Because pain interferes with respiration by limiting maximum respiratory effort, it is crucial for the parturient to be free of pain while performing spirometry so she is as close to pre-operative baseline conditions as possible. In our study, there was a maximum VAS pain score of 20 mm during coughing. This difference can also be explained by the use of a different analgesic regimen. In our study a combination of intravenous methadone (6 mg after 2 h), spinal fentanyl and rectal paracetamol was used, compared with PCA morphine (8 mg after 2 h), which was used alone in the other study [12]. We cannot exclude the possibility that mothers did not use enough PCA for fear of adverse effects on the newborn through breast-feeding.

Initiation of effective SA was associated with a decrease in spirometric parameters with the greatest decrease observed in PEFR (−30%) and MEF25−75 (−29%) values. These findings are in line with previous studies [3, 4, 12–14]. Because MEF25−75 values do not depend on patient co-operation and are comparable with PEFR values, the marked decrease observed in our study was attributable to SA and not to poor patient performance. To produce an effective cough, the patient has to inspire deeply, close the glottis, and increase the intrapulmonary pressure. Functional integrity of abdominal muscles is considered very important in cough generation [13, 18] and PEFR is a good indicator of cough effectiveness [15, 19]. During the entire study period of 3 h, PEFR values were significantly decreased in all parturients and the decrease in PEFR persisted after full resolution of SA and patient mobilization. Three hours after Caesarean section, the reduction in PEFR was still more pronounced in parturients with a BMI of > 30 kg.m−2 than in those with a normal weight (−34% vs. −17%). This is important, as vomiting in the supine position can lead to aspiration even in the conscious patient with competent laryngeal reflexes [20]. Deficiency in cough effectiveness thus adds another dimension to the high-risk profile observed in obese parturients presenting for Caesarean section. Therefore, it is imperative to implement an antacid regimen pre-operatively to reduce the risk of pulmonary acid aspiration syndrome. In our study, no case of aspiration was noted.

In contrast to the above findings that indicate an overall impairment of respiratory function following SA, we observed in a previous study that low-dose epidural analgesia during labour resulted in a small but significant improvement of respiratory function after initiation of effective epidural analgesia (VC + 7%) [21]. We hypothesised that a reduction in abdominal wall tension in the absence of intercostal muscle blockade would result in a decrease of diaphragmatic strain and thus ease breathing [21]. In the present study, however, the density and extent of motor blockade was much greater after SA using bupivacaine 0.5% than after epidural analgesia using a mixture of bupivacaine 0.125% and fentanyl. In addition, the upper sensory block level was much higher after SA than after epidural analgesia (T4 vs. T8).

We have also found that the BMI-dependency of respiratory function occurred following vertical laparotomy while evaluating different peri-operative analgesic regimens: epidural analgesia significantly improved lung volumes following surgery compared with systemic opioids, especially in the obese (unpublished observations). Therefore, obese parturients might further benefit from a combined spinal-epidural anaesthetic technique, with epidural analgesia following Caesarean section to further improve maternal pulmonary function during the immediate postoperative period.

We conclude that spinal anaesthesia in parturients scheduled for Caesarean section was associated with a BMI-dependent decrease of lung function, which persisted well into the recovery period, even longer than the actual presence of motor blockade. Sensory levels of SA did not correlate with lung volume impairment in the postoperative period. Although a direct comparison of lung volumes before and after mobilisation was not carried out, early out of bed mobilisation was most probably the reason for the significant improvement of lung volumes 3 h after the operation. This beneficial effect on the recovery of respiratory function was present in all parturients, but even more so in the obese. In normal weight parturients, the decrease in respiratory function was minor and baseline values were almost re-established 3 h after Caesarean section and mobilisation. In contrast, impairment of pulmonary function persisted in obese parturients for a longer period despite mobilisation.


The authors are indebted to the recovery room nurses for their great help. The authors also thank J. Etlinger for editorial assistance.