These conclusions can be drawn from the present study despite a potential limitation due to its nonrandomized design. Thus, according to Makuch and Simon, “if the historical-control and the experimen-tal-treatment patients are similar with regard to measured factors that are thought to be of prognostic importance, then one has a reasonable basis for employing significance tests to compare their response rates.” Given the necessity to know whether the patient groups being studied were indeed comparable (thus, at equal risk of having the outcome being assessed), every effort has been made to quantitatively assess the present study groups’ comparability.
First, demographic data and operation characteristics were similar in control and treatment groups (Table 1). The incidence of risk and high-risk patients was nearly constant, arguing against a relevant imbalance in the patients’ population characteristics. buy asthma inhalers
Second, the pretreatment general disease and sepsis severity was assessed by scoring systems in addition to relevant single parameters (Table 1). Scores are considered to be useful tools to determine whether control and treatment groups were similar and thus reduce susceptibility bias. APACHE II scoring has detected dissimilarities in illness severity between patients receiving different treatment regimens, thus preventing inappropriate conclusions regarding the effectiveness of the therapies tested. According to our results, baseline disease severity was comparable among the patient groups, and for the differences detected in a few single parameters, no impact on response or mortality could be found. Third, we evaluated the concurrent therapy as another potential source of imbalances between consecutive treatment groups. Although the number of antibiotics was higher in the Ig groups, there was no significant association between the number of antibiotics and score response and mortality rates. Furthermore, even if, due to the nonrandomized design, we cannot completely rule out a cotreatment bias, we are not aware of any further substantial change regarding the standard or supplemental treatment regimen in the Ig group compared with the historical controls.
The high-risk control group mortality (76 percent) was comparable to that of other postsurgical patients with severe sepsis (67 percent). Sepsis scores according to Elebute and Stoner in our patients (Table 1) were in the range considered indicative of a septic state as sooi} as on the first postoperative day.