The reported incidence of acute postoperative spinal infections ranges from less than 1% to 13%. The type and length of the procedure influences the infection rate. Infection rate of lumbar open discectomies without fusion is known to be 1-2%, microdiscectomy 5%, instrumented posterior lumbar fusions 2-6%, respectively5,13).

C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) have been used to diagnose postoperative infections after spinal surgery. Several studies have reported the usefulness of CRP, ESR, and WBC as a predictor of infection after spine surgery. However, it is difficult to detect infection before day 7 because incubation period where the immune system is activated to protect a living body from invading bacteria. The CRP is a too sensitive marker to distinguish the case of early infection from postoperative inflammation.

The purpose of this study was to assess the normal range and pattern of the change in plasma CRP, ESR, and WBC according to operative method and extents of spinal surgery.

Forty patients in two groups were prospectively enrolled from March of 2006 to June of 2006. Group A included 20 patients who had single level lumbar open discectomy (LOD) and group B consisted of 20 patients who had posterior lumbar interbody fusion (PLIF) with stand-alone cage after decompression. All patients were operated by the same surgeon and all did not have postoperative infection. They did not have any constitutional symptoms of infection with normal surgical wound healing and normal CRP three months after surgery.

Inclusion criteria included those 18 years-old older with surgical indication for elective surgery, either single level discectomy or PLIF and hospital length of stay of 7 days or more after the surgery. We excluded the patients who had previous spinal surgery, transfusion during the surgery, anemia, history of immunodeficiency, autoimmune disease such as rheumatoid arthritis, and any history of infection within 6 months of the surgery.

All patients had blood draws for CRP, ESR, and WBC count 1 day before surgery and postoperative day (POD) 1, POD3, and POD7. The CRP level in serum was determined with the latex agglutination methods. The normal CRP value in the authors' hospital was less than 0.3 mg/dL. ESR is a manual test in which the level of blood collected in a Western sedimentation rate tube is read by a technologist after a specific time period. Normal ESR range is less than 20 mg/dL. The WBC counts were obtained with an automatic cell counter. Preoperative and postoperative prophylactic antibiotics were administered intravenously for 7 days.

Statistical analysis was performed with using repeated measures ANOVA, and changes in effect measures over time and between groups over time were assessed. All data analysis was conducted using SAS v.9.1.

CRP, ESR, and WBC have been widely used for estimating the status of infection. Several studies have reported the usefulness of this predictor and mentioned that the combination of normal CRP, ESR, and WBC levels reliably predicts the absence of infection after elective surgery5,10,12,18).

C-reactive protein was discovered and named for its reactivity to the C polysaccharide in the cell wall of S. pneumonia, which is produced by the liver in response to inflammation, infection, malignancy, and tissue damage with relatively high sensitivity, response speed, and range compared to other acute-phase reactants1). Normal CRP concentration in healthy human serum is usually lower than 1 mg/dL, increasing slightly with aging. Higher levels are found in late pregnancy, mild inflammation, and viral infections (1-4 mg/dL); active inflammation and bacterial infection (4-20 mg/dL); severe bacterial infections and burns (>20 mg/dL)3). CRP levels normally rise within 2 to 6 hours after surgery and then go down by the third day after surgery2).

The ESR has been the measure of the quantity of RBC that precipitate in a tube in a defined time and is based upon serum protein concentrations and RBC interactions with these proteins. Inflammation causes an increase in the ESR. The ESR increases in pregnancy or rheumatoid arthritis, and decreases in polycythemia, sickle cell anemia, hereditary spherocytosis, and congestive heart failure2). Multiple factors such as patient's age, gender, RBC morphology, hemoglobin concentration, and serum levels of immunoglobulin influence the ESR. The sample must be handled appropriately and processed within a few hours to assure test accuracy. While the ESR is not a diagnostic test, it can be used to monitor disease activity and treatment response and signal that inflammatory or infectious stress is present. For example, in rheumatoid arthritis, the ESR correlates well with disease activity; however normalization of ESR often lags behind successful treatment that causes resolution of the inflammatory state.

Immediately after surgery, lymphocytes decrease due to apoptosis but neutrophils increase causing overall rise in WBC count. Therefore, the change in the peripheral WBC count, especially neutrophil count, over time serves as a useful marker of postoperative progress17).

There have been several studies that compared CRP, ESR and WBC in regards to early detection of infection. CRP has been known to be more reliable than ESR for monitoring the effectiveness of treatment for acute hematogenous osteomyelitis19), septic arthritis in children15), infected total joint arthroplasty6), and spine infection7). CRP changes more quickly than ESR and therefore CRP may be a better reflection of current inflammation. Khan et al.7) also mentioned that serial CRP level measurement may be helpful in monitoring the treatment of postoperative wound infections after spinal surgery. However, ESR does not appear to be as a sensitive marker of infection resolution.

For the explanation of prompt response of CRP, they assumed that CRP, unlike the ESR, is a fairly stable serum protein whose measurement is not time-sensitive and is not affected by other serum components. The magnitude of inflammation directly relates to the concentration of CRP20).

Numerous studies have described the normal kinetics of CRP as a rapid rise with peak value on POD2 or POD3, followed by an initial sharp fall and then a gradual decrease with normalization by POD14 to POD216,7,8,15). Larsson et al.10) analyzed the prospective study of 89 patients who underwent uncomplicated spinal surgery showed that CRP rises within the first two postoperative days and then sharply declines. ESR shows more irregular peak at POD5 and then gradually returns to normal after several weeks5,7,10,18).

Comparing with our analysis, in LOD group, CRP was normalized at postoperative 3rd day, however, in PLIF group, tendency of consistent high CRP with steady declination was detected. However, there literatures did not design the study as limit to the single procedure. In current study, we compared the changes of CRP, ESR, and WBC according to the procedures (LOD and PLIF). It is hard for us to find the literatures asserting about changes of inflammatory markers according to the procedure. For that reason, we assert that our current study is valuable for that point.

Two different surgical procedures have varying inflammatory peak responses. In spinal surgery, maximum postoperative CRP levels depend on the length and type of surgery performed9,10,14). In general, the peak response is affected by the amount of iatrogenic tissue injury at surgery. Our study shows that the peak CRP level is low in LOD group than PLIF's one.

There are a few literatures which measures and showed the normal ranges of CRP, ESR, and WBC in uninfected normal cases according to the procedures. Understanding the determinants of peak value is important because values substantially higher than expected may indicate an abnormal inflammatory response due to a complication. Deviations from normal kinetics of ESR and CRP may indicate an infectious complication. For spine surgery, reported peak values vary across studies. Kock-Jensen et al.8) reported peak CRP of 28.5 mg/L after lumbar disc surgery. Larsson et al. presented peak values for microdiscectomy (46 mg/L), open discectomy (92 mg/L), anterior lumbar fusion (70 mg/L), and posterolateral interbody fusion (173 mg/L)18). Takahashi et al.16) found significantly higher peak values after posterior lumbar instrumented fusion (91.9 mg/L) than decompression (27.8 mg/L). Deviations from normal kinetics of ESR and CRP may indicate an infectious complication.

There have been several reports about value of CRP, ESR, and WBC according to infection. Chung et al.2) reported that the postoperative ESR in uncomplicated cases remained elevated 1 week after surgery, while the CRP peaked at day 2 and normalized by day 7 after spine surgery. Mun et al.13) reported that, if there is persistent elevation or second rise around a week after surgery, the wound infection should be considered. Takahashi et al.16) found that the post-operative CRP reached its peak on day 2 and remained abnormally elevated even after six weeks. And, Mok et al.12) tried to derive the normal kinetics of CRP, opening the way for recognition of deviations that may indicate a complication. The kinetics seems to be conserved regardless of operation, magnitude, or region. A second rise or failure to decrease as expected has a high sensitivity, indicating that an infection will cause a positive result.

CRP and ESR reflect the degree of the inflammatory and surgical injuries as CRP has shown to have a higher sensitivity and specificity and is more reliable than ESR for detecting postoperative infection16). In additions, some literatures asserted that CRP is a predictable and responsive serum parameter in postoperative monitoring of inflammatory responses in patients undergoing spine surgery9,11).

Knowledge of the postoperative peak value, either by routine measurement on POD2 or 3 or estimation of the peak value based on the analysis above, allows calculation of expected values at subsequent time points because the kinetics are defined. If a patient presents with findings concerning for infection days or even weeks after surgery, CRP can be measured with a simple blood draw at that time and compared with the value expected if the patient adhered to normal kinetics12).

In current study, we excluded the patients who had transfusion during the surgery. We tried to minimize the influence of blood transfusion to CRP, WBC, ESR changes. Enright et al.4) mentioned in his prospective study of CRP concentrations pre- and post-transfusion, serum CRP levels were analyzed before and after transfusion, using a fluorescence polarization immunoassay. He found that small rises in CRP concentrations occurred following 55.6% of transfusions, however, there was no statistically significant difference between pre- and post-transfusion CRP values, and these increases also failed to reach clinical significance. He added that CRP post-transfusion of greater than 100 mg/L occurred on only one occasion, and was more likely to be due to underlying infection4). However, ESR and WBC may be affected by transfusion. In current case, we excluded the case of spinal surgery which needed transfusion, and that point made more reliable results than other literatures, reporting the CRP, ESR, and WBC with spinal surgery.

We had very critical shortage that we used the average, median value and limitation of statistical analysis due to small number of prospective study cases. However, we tried to enumerate all the measured value of this study in the table.

We presumed our prospective series might be the valuable report of comparison between surgical techniques without influence of transfusion.

We assume that normalization of CRP at postoperative 3rd day of LOD may give us relief that surgery is not infected and checking and worrying about non-normalization of CRP at postoperative 7th day of PLIF with stand-alone cage might be the unnecessary work.

For ESR, only the thing that we could find is both procedures showed steady increase until 7th postoperative day except steepness of inclination, which might mean that ESR does not help for monitoring the infection after surgery.

For WBC, we presumed that normalization of WBC at postoperative 3rd day of LOD may provide us the non-infected state of surgery, and normalization of WBC at postoperative 7th day of PLIF with stand-alone cage probably may assure of non-infection.

We tried to aware the typical pattern of CRP, ESR, and WBC in non-infected postsurgical patients and knowing the pattern may help to evaluate the early postoperative course. We found that in LOD group of non-infection, dramatic decrease of CRP was detected on POD3 and POD7 and dramatic increase of ESR on POD3 and POD7 in PLIF group of non-infection.

Both group's ESR level showed typical kinetics, not decreased until POD7 and WBC level maintain nearly equal to POD7. However, the CRP level was sharply decreased at POD3 in LOD group that of PLIF. So, we assumed that CRP would be more effective and sensitive parameter than ESR, WBC for early detection of infectious complications, especially in LOD than PLIF.