Hypotension is the most common side effect of neuraxial blocks in the obstetric patient. Over the last decades several interventions, such as pelvic tilt and the prophylactic administration of fluids or ephedrine, have been proposed to reduce the incidence of maternal hypotension. This article discusses the effectiveness of a prophylactic intravenous fluid bolus, often referred to as "prehydration", to prevent hypotension after spinal anesthesia.
The cause of hypotension after spinal anesthesia is the preganglionic sympathetic blockade resulting in vasodilatation. Arterial and venous vasodilatation reduces cardiac preload. Decreased cardiac preload in turn limits cardiac output, the main compensatory mechanism to counteract spinal vasodilatation. In the pregnant patient, compression of the vena cava by the gravid uterus further impedes venous return to the heart. If untreated, this process may lead to maternal hypotension and uterine hypoperfusion.
In 1968, Wollman and Marx (1) introduced a concept subsequently known as "prehydration" (PH). The authors proposed that intravenous volume expansion should be instituted as a prophylactic rather than therapeutic measure in response to hypotension. They reported PH to be successful in all 14 patients receiving prophylactic volume expansion after spinal anesthesia and a complete failure to prevent hypotension in five non-PH women. Although Clark et al (2) reported similar results in 1976, the remarkable success (100% prevention of hypotension) was never reproduced. Many investigators tested different PH regimens by modifying fluid characteristics (crystalloid versus colloid), volume or timing with mixed results. Repeated discussions and reports about PH led many anesthesiologists to routinely administer a fluid bolus prior to spinal anesthesia despite the lack of solid and reproducible evidence.
Graphical Representation of the Relative Risk of Hypotension
(Click graphs for larger images)
Statistics: Effect size (relative risk) is based
on the incidence of hypotension provided by the articles. The homogeneity assumption
was rejected in the crystalloid versus control study set because of outliers.
Therefore, a random effect size model was chosen.
Interpretation: The relative risk (RR) of a study refers to
the hypotension risk. For example, Riley et al found the RR to be 0.5. The risk
for hypotension in that study group was therefore twice as high if a subject
was randomized to crystalloid versus colloid prehydration.
In recent years, policy makers in medicine have called for a more scientific decision-making process. Many researchers started to critically review current practice parameters. Rout took on the concept of PH when he discovered that even large volume crystalloid PH would not decrease the incidence of postspinal hypotension (3). In a large prospective randomized trial, Rout discovered that crystalloid PH provided only a marginal benefit (4). As in many other studies, ephedrine requirements did not differ among groups. Two years later, in 1995, Jackson et al once again took on the topic of PH (5). This time crystalloid PH did not appear to offer even a small advantage.
As the body of literature on PH grew, researchers started to compare articles. The Cochran Pregnancy and Childbirth Group started a database on the evidence about prophylactic measures to prevent post-spinal hypotension (6). In their 2002 review the group concluded that no intervention reliably prevented hypotension during spinal anesthesia for cesarean section. A recent review by Frigo (7) focused on crystalloid and colloid PH and concluded that crystalloid PH was not effective.
The practicing anesthesiologist is now faced with the task of implementing the results of apparently opposite findings into clinical practice. Frustrated by the lack of consistent recommendations, many physicians choose not to disregard new evidence completely. Instead, with a little insight one might understand why research findings can vary.
With respect to post-spinal hypotension, several factors need to be considered: One can begin with the definition of the primary outcome measure, hypotension. Most articles define hypotension as a 20% decline of baseline blood pressure but many variations of this designation are used. Some use mean blood pressure, others are more generous with respect to the degree of hypotension (25% or 30% of baseline) and some do not define a baseline value at all. Another important issue is the amount of ephedrine used; obviously the severity of hypotension can be camouflaged if patients are aggressively treated with ephedrine. Other differences can be explained by variations in study design (e.g. was there a control group not receiving PH?) and statistical analysis. However, a complete discussion of the latter would go beyond the scope of this review.
A formal approach to summarizing and evaluating a multitude of randomized clinical trials was proposed in the mid-1970s when hundreds of studies of psychotherapy had produced a dizzying array of positive, null and negative results, and reviews of those studies had failed to resolve the debate. Gene Glass statistically standardized 375 psychotherapy studies, calling this method "meta-analysis" (8).
The key concept of such an analysis is based on the idea of being able to summarize and compare the results of different tests. Introducing the parameter "effect size" does this. The effect size can be based on a wide array of statistical tests used to compare outcomes in studies. Using the example of hypotension, this approach allows us to compare studies even if one reports only the incidence of hypotension using a chi-square test and another the total ephedrine requirements using the t-test.
Results are usually presented graphically displaying effect sizes of studies with their respective 95% confidence intervals. Effect sizes of included studies are then treated as dependent variables. The selection of studies should be unbiased and representative of the whole population of studies. The overall analysis of studies is based on a preliminary test, like the homogeneity test, similar to the choice of analysis in an individual study, which will be affected by items such as the normality test of the equal variance test.
Figure 1 presents the comparison of a representative sample of studies comparing (a) crystalloid preloading to control (no volume challenge) and (b) crystalloid preloading to colloid preloading (albumin, dextran or hetastarch). The analysis indicates that (a) the risk of hypotension when receiving PH is almost identical to the risk of hypotension without PH and (b) the risk of hypotension is halved when prophylactic colloid fluid expansion is compared to prophylactic crystalloid fluid expansion.
I have been interested in the concept of PH for several years and was fascinated by the results of a British survey (9). According to the authors of that survey, anesthesiologists routinely practice PH, despite a growing body of evidence challenging the concept. Today the more relevant question to be asked appears to be: How do we effect change in the practice of medicine? The necessary ingredients for a change in medical practice based on evidence appear to be: continuing education of health care providers at all levels, addressing ethical concerns and drawing modest conclusions that do not overstate the facts.
I have summarized the evidence above. Now, let me try to draw a sensible conclusion: Undisputed by most clinicians is the notion that neither crystalloid nor colloid PH reliably prevents hypotension. There appears to be little to no benefit of crystalloid PH and a distinct but small benefit of colloid PH. Based on this information, we have eliminated the nursing policy on routine crystalloid fluid administration prior to neuraxial blockade at the University of Florida. In addition we do not believe that the small benefit of colloid PH outweighs the potential risk of volume overloading and the overall substantial cost.
Michael A. Frölich, MD, DEAA