Spinal anesthesia is commonly used for elective cesarean delivery. Associated hypotension is caused by an increase in venous capacitance and a reduction in systemic vascular resistance. Because uterine blood flow is dependent on perfusion pressure, hypotension results in reduced uterine blood flow, with a potential compromise in fetal oxygenation. Maternal nausea and vomiting may also occur.
An early study reported that hypotension could be prevented with a crystalloid fluid bolus (1). Although this observation has not been replicated, fluid preload has become standard practice before regional anesthesia for cesarean delivery. Since this early work, many studies have been performed with contradictory results. Recently, the role of preload in the prevention of hypotension has been questioned (2).
This systematic review evaluates the role of central blood volume augmentation in reducing the incidence of hypotension with spinal anesthesia for cesarean delivery.
We systematically sought reports of randomized controlled trials investigating any method of increasing central blood volume to prevent hypotension during elective cesarean delivery under spinal anesthesia. Authors independently searched MEDLINE (1966–2000) and Embase (January 1988–April 2000) using medical subject (MESH) headings of “cesarean section,” “hypotension,” and “anesthesia, spinal.” “Leg wrapping” and “trendelenberg” were also used as text terms to capture other methods of increasing preload. The last search was performed on May 1, 2000. The Cochrane Library (Issue 1, 2000) was also searched. Additional reports from reference lists of retrieved and review articles, hand searching of non-MEDLINE journals, and abstracts of major anesthesia meetings (1995–2000) were used to locate subsequent publications. No abstracts, correspondences, or unpublished observations were included.
Quality of the Trials
At least two authors assigned a quality score to each report using a three-item, validated scoring system (3). We gave one point to studies described as randomized and an additional point if randomization was described and appropriate (i.e., randomization table, random number generator). A randomization point was deducted if an inappropriate method was used (chart number, day of the week, alternate allocation). One point was given if the study was blinded and an additional point if the blinding was appropriate. Finally, a point was awarded if all enrolled patients were accounted for, either by showing to which group(s) they were allocated or by describing why they were withdrawn from the study. Thus, a total of five points were possible. Any disagreements in scoring were resolved by consensus.
The primary outcome was the incidence of hypotension, as defined by the authors of the original studies. Patients receiving no volume expansion were the Control group. In some cases, a small volume of fluid was compared with a large volume; in that case, the former was considered the Control group. Secondary outcomes included maternal nausea, the use of vasopressors, neonatal Apgar scores, and umbilical cord pH.
We intended to combine studies statistically and stratify results into three groups: 1) crystalloid preloading, 2) colloid versus crystalloid, and 3) mechanical methods of increasing autotransfusion. This was not done because: 1) in general, quality of the studies was poor, 2) outcome measures were not uniform across studies, 3) control groups were not comparable to each other (for example, some of the control groups also received a fluid bolus) (4–7), 4) different local anesthetics and doses were used, and 5) some studies used prophylactic vasopressors while others did not. Therefore qualitative methods and tables summarize the results. Also, l’Abbé scatter plots were calculated for each method of increasing central blood volume. L’Abbé plots, tools used to graphically represent information found in a number of trials, assist the reader in understanding the collected data (8). Each point on the scatter plot represents one trial with the proportion of patients receiving a control treatment plotted against the proportion receiving an experimental treatment.
The MEDLINE search yielded 91 articles in English, of which 23 identified randomized controlled trials that met our inclusion criteria. They occurred in three categories: comparisons of large versus small dose of crystalloid, colloid versus crystalloid or different colloid regimens, and mechanical or patient positioning as a means of increasing preload. The studies include 1504 patients, with individual studies ranging from 16–160 patients. (Tables 1–3)
One article was not, by strict definition, a randomized trial (9). Although the first 40 patients were randomized in blocks of 2, subsequent patients were enrolled by using a “play the winner” strategy, and accounted for more than half the patients. We chose to include this study, but a quality score was not assigned.
Prevention of Hypotension Using Crystalloid Preload Before Spinal Anesthesia
We retrieved 9 articles comparing two crystalloid volumes. (Table 1) (1,4–7,9–12). One article was divided into two groups, Group 1 having elective cesarean delivery and Group 2 having cesarean delivery after labor (10). Neither group received prophylactic left uterine displacement. All remaining studies involved healthy women at full-term gestation undergoing elective cesarean delivery. In total, 533 patients were involved, 312 in the Experimental group (larger preload volume) and 221 in the Control group (smaller preload volume). Only three articles had a quality score > 2 (7,11,12); one was not assigned a quality score (9), and the earliest study had a quality score of 0 (1).
In all studies, a balanced salt solution was used, and in two studies, the solution contained dextrose (1,10). Total volume of the administered preload varied widely, ranging from 0–1787 mL in the Control group to 997–2970 mL in the Experimental group.
There was a significant effect on the incidence of hypotension in patients receiving a larger preload volume in three studies (1,9,10). The remaining studies demonstrated no significant difference (4–7,11,12). Three of the studies showed a trend toward a decreased incidence of hypotension in the treated group (6,7,12), whereas the other three studies (4,5,11) had similar incidences of hypotension in both groups.
The use of prophylactic and therapeutic ephedrine was different between studies and is outlined in Table 1. The l’Abbé plot for crystalloid preload is depicted in Figure 1. Because Olsen’s study did not report the incidence of hypotension, it could not be included in the graph.
Of the five studies reporting umbilical cord pH or [H+](1,6,9,11,12), only one (1) demonstrated a significant difference. The incidence of Apgar scores < 7 did not differ with the exception of the article by Marx et al. (1). Only two studies addressed the issue of maternal nausea and vomiting (4,12), and neither reported a difference.
Prevention of Hypotension Before Spinal Anesthesia Using Colloid
Seven studies compared crystalloid and colloid. (Table 2) (13–19). Six had a crystalloid control (13–16,18,19), and one (17) compared different colloid types and regimens. Some investigators used equal volumes of crystalloid and colloid, whereas others used larger volumes of crystalloid to account for redistribution from the intravascular compartment. Only three studies attempted to blind the investigators to the type of solution given (15,18,19). Sample sizes varied between 26 and 160 patients with a total of 559 parturients.
Mathru et al. (13) used albumin whereas later studies used either hetastarch (HES), pentastarch, modified gelatin, or dextran. Five studies showed a significant decrease in the incidence of hypotension comparing crystalloid to colloid (13,15,16,18,19). The difference in incidence of hypotension in Karinen et al.’s (14) study was similar to that of other investigators but was not statistically significant because of low power. Vercauteren et al. (17) found less hypotension with HES compared with modified gelatin.
Six studies examined neonatal outcomes by using Apgar scores or umbilical artery pH (13–15,17–19). Only Mathru et al. (13) showed improved neonatal outcome in the colloid group. All studies had protocols to rapidly treat hypotension with ephedrine. No difference in nausea and vomiting was noted in the two studies that addressed this issue.
The l’Abbé plot for colloid preload is depicted in Figure 2. Vercauteren’s study did not compare crystalloid to colloid and is therefore not included.
Prevention of Hypotension Before Spinal Anesthesia Using Mechanical Interventions
Seven studies compared the incidence of hypotension with autotransfusion from the lower limbs to a control group with no intervention, achieved by wrapping the legs (20–22), leg elevation (20–23), tilting the bed head down (24), thromboembolic stockings (25), or inflatable splints or boots (23,26) (Table 3). Some patients received more than one intervention (20–23).
Most studies were small (range, 24–97) with two to four groups. Quality scores were poor, with only one score > 2. All patients received between 500 and 2000 mL of prophylactic crystalloid, and none received prophylactic ephedrine. Except for one study (23), patients had left uterine displacement ensured. There were 447 patients in the Control group and 339 patients in the Experimental group.
Wrapping the legs consistently and significantly reduced the incidence of hypotension compared with control or leg elevation. (Table 3) (20–22) Tilting the patient head down was ineffective (24). Inflatable splints reduced the incidence of hypotension from 83% to 48% after a 15-mL/kg crystalloid preload (26). This was statistically significant but the authors concluded that the incidence of this complication was still unacceptably frequent. The use of inflatable boots did not reduce the incidence of hypotension but reduced the need for vasopressors and delayed the onset of hypotension (23) (Table 3).
There was no difference in the incidence of adverse outcomes between groups in studies that measured cord pH or Apgar scores. There was a significant reduction in maternal nausea when inflatable splints were used (26).
The l’Abbé plot for mechanical interventions is depicted in Figure 3.
Prevention of Hypotension Using Crystalloid Preload Before Spinal Anesthesia
Marx et al. (1) demonstrated a 0% incidence of hypotension when patients received a prophylactic infusion of 1 L of lactated Ringer’s solution with 5% dextrose within 30 minutes before the administration of spinal anesthesia (n = 18). However, patients were not appropriately randomized because they were offered a choice of spinal or general anesthesia. No information regarding blinding of the data collector or anesthesiologist was included. Subsequent investigators have been unable to reproduce these results.
Only one article, involving the largest number of patients, addressed the issue of a power analysis in the study design (9). Because of ethical concerns of not administering a preload, the authors chose a statistical model incorporating a sequential design using a “play the winner” rule allowing the study to be terminated if a large difference in treatment effect was noted (9). The study, designed to detect a 20% difference in the incidence of hypotension between groups, assumed an approximate incidence of 50% in the preload group with statistical significance at the 10% level (α = 0.1), one-tailed and 90% power (9). The study was terminated after the third statistical analysis with 78 patients in the preload and 62 in the Nonpreload group. The authors reported a statistically significant decreased incidence of hypotension in the Preload group but commented that the administration of a fluid bolus did not eliminate the risk of hypotension and they did not advocate routine fluid loading. However, more patients were entered into the Experimental or larger preload group, indicating that the “winners” were the patients who received a preload. This adds additional support to the notion that volume preloading made a difference in the incidence of hypotension.
Another limitation of the studies listed was the wide variation in preload volume in the control and experimental groups. In five studies, the control group received little or no preload (1,9–12). In contrast, “control” (Nonpreload group) received preload volumes ranging from 750 to 1800 mL in four other studies (4–7).
Criteria for vasopressor administration were not standardized, but in most cases, ephedrine was given in bolus doses to maintain blood pressure at 80%–90% of baseline values. In two studies, investigators gave prophylactic ephedrine to the Control or Nonpreload group (6,11). Jackson et al. (11) began an infusion of 500 mL of Hartmann’s solution with 60 mg of added ephedrine at an unspecified rate when spinal medications were given. The Experimental group received ephedrine as deemed necessary by the anesthesiologist. Olsen et al. (6) administered a bolus dose of ephedrine 0.15 mg/kg IV followed by an infusion of ephedrine 0.4 mg/kg to patients who did not receive a preload. Husaini and Russell (12) used prophylactic ephedrine infusions for both groups, and either an increase in the infusion rate or bolus doses of ephedrine were used to treat hypotension and to maintain blood pressure in the “normal” range. In summary, differences in study protocols account for the heterogeneity of the results reported. (Fig. 1) Crystalloid loading will not reliably prevent hypotension.
Prevention of Hypotension Using Colloid Preload Before Spinal Anesthesia
These studies indicate that colloid is superior to crystalloid in preventing postspinal hypotension for elective cesarean delivery. Only the study with the smallest sample size showed no difference (14). The study by Vercauteren et al. (17) did not have a crystalloid control but was included in this section because two volumes of colloid were compared.
A limitation of these results is the difference in the volumes chosen by the investigators. Three investigators chose a ratio of 2:1 for crystalloid to colloid to account for the redistribution of crystalloid out of the intravascular space (14,15,18), whereas two chose to give the same volume (13,19) and one performed a dose ranging study with ratios of 2:1 and 3:1 (16). Different definitions of hypotension and ephedrine protocols were used, but all studies rapidly treated hypotension when it occurred.
A physiologic explanation of the differences between crystalloid and colloid can be found in Ueyama et al.’s (16) study which elegantly showed that, at 30 minutes, only 28% of the administered lactated Ringer’s solution remained in the intravascular space compared with 100% of the HES solution. With increasing volumes of HES, there was an increase in blood volume and cardiac output and a decrease in the incidence of hypotension. Colloid is a more effective volume expander because it remains longer in the intravascular compartment. Nevertheless, even aggressive expansion of the intravascular space cannot completely abolish the risk of hypotension.
Rout and Rocke (2) suggest that colloids may contribute to decreased afterload via atrial natriuretic peptide release from a stretched atrium or by peripheral vasodilation caused by reduced arterial blood oxygen content. Reducing before central sympathetic block may play an important role in the prevention of postspinal hypotension.
Excessive amounts of intravascular fluid may cause volume overload and pulmonary edema. Karinen et al. (14) attempted to measure central volume by monitoring central venous pressure. They noted a significant increase in central venous pressure in both groups after preloading, a return toward normal after the induction of spinal anesthesia, followed by a steady increase until delivery and no change thereafter. There were no differences between 1 L of crystalloid and 500 mL of colloid. There were no reports of pulmonary edema in any study.
Various colloids were used and Vercauteren et al. (17) showed that HES was superior to modified gelatin but failed to offer an explanation. All of the artificial colloids carry a risk of anaphylactoid reactions although no cases were reported (27). Ring and Messmer (28) reported the incidence of allergic reactions to colloids to be: human serum albumin, 0.011%; gelatin, 0.115%; dextran, 0.32%; and hydroxyethyl starch, 0.085%. Pentastarch, currently available in Europe and Canada, but not in the United States, has the smallest risk (29). Albumin, as a human blood product, carries a theoretical risk of disease transmission.
Overall, colloids can be effective at reducing the incidence of hypotension, with no significant maternal or neonatal side effects. The l’Abbé plot shows a cluster of trials to the left of the equality line, suggesting the usefulness of the intervention. This benefit must be weighed against the increased cost of the colloid and the incidence of adverse reactions (30).
Prevention of Hypotension Using Mechanical Means
Leg wrapping, by using an Esmarch bandage from the foot or ankle to mid-thigh, was the most effective means of reducing the incidence of hypotension compared with the control groups (no intervention or leg elevation alone). In each case, the incidence of this complication, as defined by the authors was reduced to less than 20%(20–22). The use of antithromboembolic stockings was also effective (25). No study demonstrated a reduction in the amount of ephedrine used in each group or any difference in neonatal outcome.
In contrast to other methods of leg compression, inflatable boots or splints seemed to be less effective (23,26). Both interventions describe compression from the foot to the upper thigh. The reason for the discrepancy is not apparent, although these methods may be less effective in increasing central blood volume compared with leg wrapping or antithromboembolic stockings.
There may be several factors accounting for the success of leg wrapping in increasing central blood volume at cesarean delivery. First, there is approximately 150 mL of blood in the legs of nonpregnant subjects without spinal anesthesia (31). During pregnancy, venous blood volume increases in the lower extremities, particularly after the 30th week of gestation, with spinal anesthesia further increasing the volume of blood in the legs by induction of a sympathectomy (32,33). The effect of leg wrapping on the central volume has not yet been measured in parturients. The l’Abbé plot shows a cluster of trials to the left of the equality line, which suggests that mechanical means of increasing central blood volume appears better than no intervention in the prevention of hypotension with spinal anesthesia.
The common clinical practice of volume expansion with crystalloid is not uniformly effective in reducing the incidence of maternal hypotension after spinal anesthesia for cesarean delivery. Colloid administration is more consistent but is associated with additional risks and costs. Mechanical means of central blood volume expansions also appear to be effective but have not come into widespread use. Due to common differences in study design, it is impossible to determine whether prophylactic volume administration affects therapeutic vasopressor use.
In all studies, hypotension of short duration had no adverse effect on the neonate in normal term pregnancies. However, maternal side effects were not measured. This deficiency should be addressed in future studies.
1. Marx G, Cosmi E, Wollman S. Biochemical status and clinical condition of mother and infant at Cesarean section. Anesth Analg 1969; 48: 986–94.
2. Rout C, Rocke DA. Spinal hypotension associated with cesarean section: will preload ever work? Anesthesiology 1999; 91: 1565–7.
3. Jadad A, Moore R, Carroll D. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996; 17: 1–12.
4. Kangas-Saarela T, Hollmén A, Tolonen U, et al. Does ephedrine influence newborn neurobehavioural responses and spectral EEG when used to prevent maternal hypotension during caesarean section? Acta Anaesthesiol Scand 1990; 34: 8–16.
5. Rout C, Akoojee S, Rocke D, Gouws E. Rapid administration of crystalloid preload does not decrease the incidence of hypotension after spinal anaesthesia for elective caesarean section. Br J Anaesth 1992; 68: 394–7.
6. Olsen K, Feiblerg V, Hansen C, et al. Prevention of hypotension during spinal anaesthesia for caesarean section. Int J Obstet Anesth 1994; 3: 20–4.
7. Park G, Hauch M, Curlin F, et al. The effects of varying volumes of crystalloid administration before cesarean delivery on maternal hemodynamics and colloid osmotic pressure. Anesth Analg 1996; 83: 299–303.
8. L’Abbé KA, Detsky AS, O’Rourke K. Meta-analysis in clinical research. Ann Int Med 1987; 107: 224–33.
9. Rout C, Rocke D, Levin J, et al. A reeevaluation of the role of crystalloid preload in the prevention of hypotension associated with spinal anesthesia for elective cesarean section. Anesthesiology 1993; 79: 262–9.
10. Clark R, Thompson D, Thompson C. Prevention of spinal hypotension associated with Cesarean section. Anesthesiology 1976; 45: 670–4.
11. Jackson R, Reid J, Thorburn J. Volume preloading is not essential to prevent spinal-induced hypotension at Caesarean section. Br J Anaesth 1995; 75: 262–5.
12. Husaini S, Russell I. Volume preload: lack of effect in the prevention of spinal-induced hypotension at caesarean section. Int J Obstet Anesth 1998; 7: 76–81.
13. Mathru M, Rao T, Kartha R, et al. Intravenous albumin administration for prevention of spinal hypotension during Cesarean section. Anesth Analg 1980; 59: 655–8.
14. Karinen J, Rasanen J, Alahuhta S, et al. Effect of crystalloid and colloid preloading on uteroplacental and maternal haemodynamic state during spinal anaesthesia for caesarean section. Br J Anaesth 1995; 75: 531–5.
15. Riley E, Cohen S, Rubenstein A, Flanagan B. Prevention of hypotension after spinal anesthesia for cesarean section: six percent hetastarch versus lactated Ringer’s solution. Anesth Analg 1995; 81: 838–42.
16. Ueyama H, Le H, Tanigami H, et al. Effects of crystalloid and colloid preload on blood volume in the parturient undergoing spinal anesthesia for elective cesarean section. Anesthesiology 1999; 91: 1571–6.
17. Vercauteren M, Hoffmann V, Steenberge AV, Adriaensen H. Hydroxyethylstarch compared with modified gelatin as volume preload before spinal anaesthesia for Caesarean section. Br J Anaesth 1996; 76: 731–3.
18. Lin C-S, Lin T-Y, Huang C-H, et al. Prevention of hypotension after spinal anesthesia for Cesarean section: dextran 40 versus lactated ringer’s solution. Acta Anaesthesiol Sin 1999; 37: 55–9.
19. French GWG, White JB, Howell SJ, Popat M. Comparison of pentastarch and Hartmann’s solution for volume preloading in spinal anaesthesia for elective Caesarean section. Br J Anaesth 1999; 83: 475–7.
20. Bhagwanjee S, Rocke DA, Rout CC, et al. Prevention of hypotension following spinal anaesthesia for elective caesarean section by wrapping of the legs. Br J Anaesth 1990; 65: 819–22.
21. van Bogaert LJ. Prevention of post-spinal hypotension at elective cesarean section by wrapping of the lower limbs. Int J Gynecol Obstet 1998; 61: 233–8.
22. Rout C, Rocke D, Gouws E. Leg elevation and wrapping in the prevention of hypotension following spinal anaesthesia for elective Caesarean section. Anaesthesia 1993; 48: 304–8.
23. James FM, Greiss FCJ. The use of inflatable boots to prevent hypotension during spinal anesthesia for cesarean section. Anesth Analg 1973; 52: 246–51.
24. Miyabe M, Sato S. The effect of head down tilt position on arterial blood pressure after spinal anesthesia for cesarean delivery. Reg Anesth 1997; 22: 239–42.
25. Sood PK, Cooper PJF, Michel MZ, et al. Thromboembolic deterrent stockings fail to prevent hypotension associated with spinal anaesthesia for elective caesarean section. Int J Obstet Anesth 1996; 5: 172–5.
26. Goudie TA, Winter AW, Ferguson DJM. Lower limb compression using inflatable splints to prevent hypotension during spinal anaesthesia for caesarean section. Acta Anaesthesiol Scand 1988; 32: 541–4.
27. Morgan PJ. Allergic reactions. In: Birnbach D, Gatt S, Datta S, eds. Textbook of obstetric anesthesia. New York: Churchill Livingstone, 2000: 475–86.
28. Ring J, Messmer K. Incidence and severity of anaphylactoid reactions to colloid volume substitutes. Lancet 1977; 1: 466–9.
29. McHugh GJ. Anaphylactoid reaction to pentastarch. Can J Anaesth 1998; 45: 270–2.
30. Weeks S. Reflections on hypotension during Cesarean section under spinal anesthesia: do we need to use colloid? Can J Anaesth 2000; 47: 607–10.
31. Rutlen DL, Wackers FJ, Zaret BL. Radionuclide assessment of peripheral intravascular capacity: a technique to measure intravascular volume change. Circulation 1981; 64: 146–52.
32. Barwin BN, Roddie IC. Venous distensibility during pregnancy determined by graded venous congestion. Am J Obstet Gynecol 1976; 125: 921–3.
33. Green NM, Brull SJ. Physiology of spinal anesthesia. 4th ed. Baltimore: Williams and Wilkins, 1993.