Secondary Logo

Journal Logo


A reliable diagnostic method of surgical site infection after posterior lumbar surgery based on serial C-reactive protein

Kim, Min Hyung MD, PhDa; Park, Jong-Hyeok MD, PhDb,; Kim, Jong Tae MD, PhDb

Author Information
International Journal of Surgery: Global Health: September 2021 - Volume 4 - Issue 5 - p e61
doi: 10.1097/GH9.0000000000000061
  • Open


Surgical site infection (SSI) is one of the serious complications associated with spinal surgery. The incidence of SSI is relatively frequent, and is reported to range from 0.7% to 11%1,2. SSI requires long-term antibiotic therapy, prolonged hospital stay, and surgical debridement or revision surgery3,4. Nevertheless, the diagnosis of SSI is a clinical challenge, and is difficult to distinguish from the normal postoperative course. Traditional indicators of infection such as fever, leukocytosis, pain, and delirium are generally encountered during acute postoperative phase5. The signs and symptoms of SSI are also ambiguous and vague in the early phase.

Therefore, laboratory tests such as leukocyte count, segmented neutrophils, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) are commonly used for the detection of SSI. These tests are generally elevated in the presence of systemic inflammation except in cases involving specific underlying diseases such as chronic renal or hepatic disease6. Measurement of CRP is popular because it is a quantitative test with predictable kinetics. It is more reliable to monitor postoperative infection based on comparison with ESR5. The normal kinetics of CRP relative to surgery is characterized by rapid production for two to three days after surgery until the peak level is reached, followed by reduction and eventual return to normal range7,8. According to the normal kinetics of CRP, there were several studies that appropriate interpretation or clinical application of serial changes in laboratory tests after spinal surgery2,6,9. However, the elevation of CRP frequently occurred in postoperative patients. Many postoperative patients showed elevated CRP level during admission.

The goal of the present study is to evaluate serial postoperative laboratory tests and parameters for the early detection of SSI. The authors also describe the limitations of laboratory tests for early detection of SSI. Finally, we discuss the validity of serial CRP measurements as an indicator of SSI for clinical application.

Material and methods

Patient selection

Between January 2018 and December 2019, data of consecutive patients who underwent posterior lumbar decompression or fusion surgery were reviewed retrospectively by 2 neurospine surgeons (J.T.K., J.-H.P.). Institutional review board approval was obtained (2021-0790-0001). Cervical and thoracic lesions were excluded due to less extent of dissection than lumbar lesion. Patients with primary spondylitis, autoimmune disease, immunosuppressive treatment, liver disease, malignancy, and chronic renal disease were excluded. The data including characteristics of patients, underlying disease, type of surgery, the number of spinal levels treated, and diabetes mellitus (DM) were collected. All patients were injected with prophylactic antibiotics intravenously within one hour of incision. The antibiotics consisted of cefazolin 1 g every 8 hours or in patients with penicillin allergy, vancomycin 1 g every 12 hours, and continued for 24 hours after surgery or until postoperative drains were removed, usually on postoperative day (POD) 1 (POD1).

Definition of SSI

The diagnosis of SSI was defined according to the Center for Disease Control and Prevention Criteria for superficial and deep incisional SSI10. The diagnosis of SSI was based on at least one of the following criteria: purulent discharge from superficial and deep wounds, organisms isolated from an aseptically obtained culture of fluid and tissue at the incision site. SSI was characterized by at least one of the following signs or symptoms: back pain or tenderness at surgical site, localized swelling or redness, and fever. In addition, we diagnosed SSI when postoperative infection was confirmed in magnetic resonance image (MRI) with contrast enhanced by radiologist. A diagnosis of SSI was followed by prompt treatment with empirical antibiotics. Subsequently, the isolated organisms were identified. The antibiotics were changed according to the sensitivity of the bacteria. If necessary, irrigation or reoperation was performed to remove epidural abscess.

When patients without pain at surgical site showed fever and elevated laboratory parameters, we evaluated secondary causes of infection such as urinary tract infection, pneumonia, and atelectasis.

Laboratory tests

Laboratory tests including leukocyte count, segmented neutrophil, ESR, and CRP measurements were routinely performed preoperatively and on POD1, POD3, POD7, POD14, and POD30 until laboratory parameters were normalized. Normal limits of leukocyte count, segmented neutrophil, ESR, and CRP were defined as 4–9.9×103, 39%–72%, 0–20 mm/h, and 0–5 mg/L, respectively. CRP was measured using the latex agglutination method. Based on the previous kinetics studies of postoperative CRP values, the peak CRP value was detected on POD2 or POD3, followed by rapid decline until POD7, and gradual decline thereafter8,11. We defined the 2 standard groups based on the increase and decrease of CRP. In general, it was considered that SSI would be higher in the group with increased CRP, so patients were divided into two group based on the increase or decrease of CRP on POD7. Increasing CRP (ICRP) group shows a higher CRP value on POD7 than on POD3, whereas decreasing CRP (DCRP) group showed a lower value of CRP on POD7 than on POD3. Although the normal limit of CRP is known, the value of CRP differed according to the surgical type, level, and patients’ sex. We also calculated the change rate in CRP as follows:ThechangerateofCRP=CurrentvalueofCRPPreviousvalueofCRPPreviousvalueofCRP×100.

Statistical analysis

We used paired and Student t tests and 1-way analysis of variance to compare laboratory values between ICRP and DCRP groups. Receiver operating characteristic (ROC) analyses were used to estimate the optimized cut-off values of laboratory tests for the detection of SSI. The cut-off values of laboratory tests were evaluated using the area under curve (AUC) of the ROC plots. P values <0.05 were considered statistically significant. SPSS version 20.0 (SPSS Inc., Chicago, IL) was used for the statistical analysis.


Clinical information

Overall, 189 patients (91 men, 98 women) who met the inclusion criteria were enrolled in this study. The mean age was 64.5±12.5 years. One hundred nineteen patients were treated with instrumentation (63%). The distribution of each surgical level was: 91 in 1, 76 in 2, 15 in 3, and 7 in 4 levels. Revision surgery accounted for 33 (17.5%) of 189 patients. Patients with DM were 54 (28.6%). Clinical information is summarized in Table 1.

Table 1 - Clinical demographics of patients (n=189).
Variables Value
Mean age (y) 64.5±12.5
 Male 91
 Female 98
Surgical level
 1 91
 2 76
 3 15
 4 7
Surgical type
 Decompression 70
 Fusion 119
Revision surgery, n (%) 33 (17.5)
Diabetes mellitus, n (%) 54 (28.6)
Values are mean±SD.

Characteristics of SSI

SSIs were found in 17 (8.8%) patients. The incidence of SSI differed according to age: 11.3% in patients aged 65 years and above, which was higher than 6.0% in patients with below 65 years. SSI also occurred in 2 patients (2.9%) undergoing decompression surgery and in 15 patients (12.6%) treated with instrumental surgery. It was observed about 4-fold more frequently in instrumental surgery compared with decompression. SSI was frequently associated with increasing surgical level: 6.6% with 1-level surgery, 10.5% with 2-level surgery, 13.3% with 3-level surgery, and 14.3% following 4-level surgery. The previous history of surgery was significantly associated with SSI, and 21.2% of cases undergoing revision surgery were diagnosed with SSI. SSIs accounted for 13.3% of patients with DM compared with 7.4% in patients without DM. Among these variables, instrumental and revision surgery was significantly associated with SSI in χ2 analysis (P<0.03, <0.02). These results are summarized in Table 2.

Table 2 - Characteristics of surgical site infection.
Variables Incidence of Surgical Site Infection (%) P
Age (y) 0.157
 ≥65 11.3
 <65 6.0
Sex 0.436
 Male 9.9
 Female 8.2
Surgical type 0.018*
 Decompression 2.9
 Fusion 12.6
Surgical level 0.695
 1 6.6
 2 10.5
 3 13.3
 4 14.3
Revision surgery 21.2 0.015*
Diabetes mellitus 13 0.176
*Statistical significance (P<0.05).

Staphylococcus species were found in 10 of 17 cases, the most common cause of SSI after posterior lumbar surgery (Table 3). Streptococcus viridian, Enterococcus faecium, Klebsiella pneumonia, and Enterobacter cloacae were also isolated. No organism was detected in 3 patients finally. We performed surgical irrigation in 2 patients with SSI. Four patients were finally surgically retreated by exchanging or removing loose screws, and occasionally via reimplantation.

Table 3 - Serial CRP measurements and clinical information of 17 patients with surgical site infection.
Case No. Preop Postop POD3 POD7 POD14 POD30 POD60 Operation Signs Symptoms Culture
1 0.8 0.4 36 16.3 16.9 1.6 1D F AP, RP Staphylococcus
2 1.5 0.3 20.8 101 1.9 0.6 0.4 1D F, D AP, RP Staphylococcus
3 2.7 2.5 5.7 188 125.5 43.4 16.1 3F F, D, R AP, RP Staphylococcus
4 1.3 0.7 125 49 61.6 12.1 2F F AP Staphylococcus
5 0.7 0.5 155.8 24.1 13.7 9 1F F AP No growth
6 14.5 15.6 157.4 158.3 100 120 3F F AP No growth
7 32.7 28.8 45.1 21.4 19.2 38 4F F AP Enterobacter
8 0.5 0.9 21.4 8.9 19.7 83.8 2F F, R AP Staphylococcus
9 0.5 0.4 63.2 11 71.1 25.5 2F F AP Streptococcus
10 2.2 2.2 144.2 47 7.4 5.1 1F F AP Staphylococcus
11 59.8 9.2 42.4 167.4 154.7 135.7 22.5 2F F AP Staphylococcus
12 2.7 2.4 105.2 66.9 38.4 67.8 1F F AP Staphylococcus
13 4.7 5.4 11.6 16.3 77.7 4.1 2F F, R AP Enterococcus
14 2.2 1.8 82 61 72.8 47.8 3.2 2F F AP, RP No growth
15 1 0.5 0.9 0.5 0.4 25.71 0.39 1F F AP Staphylococcus
16 1 0.9 75.2 130.5 66.6 15.1 1F F, R AP, RP Staphylococcus
17 8.9 5.1 143.9 76.1 46.3 7.6 2F F AP, RP Klebsiella
1D indicates 1-level decompression; 1F, 1-level fusion; 2F, 2-level fusion; 3F, 3-level fusion; 4F, 4-level fusion; AP, axial pain; CRP, C-reactive protein; D, discharge; F, fever; POD, postoperative day; R, redness; RP, radiating pain; SSI, surgical site infection.

Analysis of laboratory tests

The results of serial laboratory tests ICRP and DCRP groups are summarized in Figure 1. Both groups showed similar pattern of graph of laboratory test except CRP. Mean leukocyte counts on POD7 were significant between 2 group that were 9.0±3.8×103 in ICRP and 7.3±2.2×103 in DCRP group, but otherwise were not significant. Mean segmented neutrophils were not significant during POD30. Mean ESR in ICRP group was observed higher than that in DCRP group at POD7 and 14 but, mean ESR showed similar pattern in both groups. Mean CRP was observed delayed peak value at POD7 in ICRP unlike DCRP group, but there was not significant difference of CRP value at POD14 in both group. The average ESR was higher than normal limit in both groups on POD7. Once the values of ESR exceeded the upper limit of normal value of ESR, it failed to return to the normal value of ESR until POD30. Therefore, it was not appropriate to detect SSI based on ESR. According to the graphic patterns presented in Figure 1, the normal values of leukocyte count, segmented neutrophil, and ESR were not useful in detecting SSI. The average value of CRP both groups exceeded the upper limit of normal value of CRP after POD3. Therefore, application of the normal value of CRP to detect SSI was limited. Sixteen patients with SSI except 1 showed higher CRP than the normal value (5 mg/L), but 100 patients on POD7 and 81 patients on POD14 both groups also showed elevated CRP levels more than normal value.

Figure 1:
Mean values of laboratory tests in increasing (ICRP) and decreasing CRP (DCRP) groups: leukocyte count (A), segmented neutrophil (B), erythrocyte sedimentation rate (C), and CRP (D) in ICRP and DCRP groups. CRP indicates C-reactive protein; POD, postoperative day.

We summarized SSI and secondary infection in Figure 2. Forty-three patients represented increasing CRP on POD7 were classified to ICRP group, it accounted for 22.8%. Only 6 patients developed SSIs among ICRP group. We also found 8 patients in ICRP group represented secondary infections such as urinary tract infection10, atelectasis6, pseudomembranous colitis3, deep vein thrombosis3, and pneumonia3. However, no signs of infection were detected in the other 29 patients. These patients showed increased CRP without any infection. Eleven SSIs were also observed in DCRP group. Twenty-six patients in DCRP group showed a secondary elevation on POD14 and 5 of 26 patients developed SSI. Five patients were also found to carry secondary infections such as urinary tract infection3, atelectasis3, pseudomembranous colitis3, pulmonary embolism3, and pneumonia3. However, the remaining patients showing secondary elevation revealed no causes of infection. Even 6 patients without secondary elevation also developed SSI. They showed a slow decrease in CRP level on POD7 and POD14.

Figure 2:
Distribution of surgical site infection and secondary infection in increased CRP (ICRP) group and decrease CRP (DCRP) groups based on laboratory parameters. CRP indicates C-reactive protein; POD, postoperative day; SSI, surgical site infection.

Our results were different to the normal kinetics of CRP in postoperative patients6,12,13. Even 42.9% of patients showed elevated CRP more than the normal value on POD14. A continuous increase or a decrease failure of CRP after the peak time was observed in patients with SSI. We used the change rate of CRP to detect SSI due to this feature of CRP instead of the normal value of CRP.

In ICRP group, the mean change rate of CRP between POD3 and POD7 was 202.8±529.2%. The cut-off value of CRP variation for SSI detection was more than 40.3%, and the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 83.3%, 56.9%, 23.9%, and 95.4%, respectively, based on the ROC curve. The mean change rate of CRP between POD7 and POD14 was −42.1±108.3%, and the cut-off value of CRP variation for SSI detection was more than −49.6%. The sensitivity, specificity, PPV, and NPV were 83.3%, 82.4%, 43.5%, and 96.8%, respectively (Fig. 3). On POD14, the accuracy of method using CRP was 82.5% in ICRP group. Moreover, in DCRP group, the mean change rate of CRP between POD3 and POD7 was −61.5±22.5%, and the cut-off value of CRP variation to detect SSI was more than −67.6%. The sensitivity, specificity, PPV, and NPV were 81.8%, 50.4%, 11.8%, and 97.2%, respectively, based on ROC curve. The mean change rate of CRP between POD7 and POD14 was 118.3±1115.3%, and the cut-off value of CRP change rate to detect SSI was more than −43.3%. The sensitivity, specificity, PPV, and NPV were 90.9%, 68.1%,18.8%, and 98.9%, respectively (Fig. 3). On POD14, the detection of SSI is 69.8% in DCRP group.

Figure 3:
Receiver operating characteristic curves representing the changes in C-reactive protein (CRP) level for detection of surgical site infection. A, Change in CRP levels between postoperative day (POD)3 and POD7 (a solid line) and between POD7 and POD14 (a dotted line) in the increasing CRP group. B, Change in CRP levels between POD3 and POD7 (a solid line) and between POD7 and POD14 (a dotted line) in the decreasing CRP group.


The earlier SSI is diagnosed and treated, the better it are the clinical outcomes14,15. Therefore, early detection of SSI is important for the development of a surgical strategy. However, it is challenging in that clinical examination for SSI is vague and ambiguous due to the incubation period and is dependent on the patients’ expression16. Therefore, many surgeon has tried to find objective and reliable methods such as laboratory tests to diagnose SSI, but there are still controversy17.

Similar to previous studies, leukocyte count, segmented neutrophil, and ESR were not reliable parameters to detect SSI18. Numerous studies have described the normal kinetics of CRP9,13,19. A peak value of CRP was observed on POD2 or 3 followed by an initial sharp decline and then a gradual decrease with normalization by POD14 to POD21. A similar CRP kinetics was found in 60.3% of patients included in the study. However, 22.8% of all the patients showed a continuous increase in CRP on POD7 higher than the peak value observed on POD3. We defined these patients as ICRP group. SSIs or other secondary infections were established in 58.1% of the ICRP group, whereas 41.9% of the ICRP group showed a CRP increase without any infectious cause. Fugita et al16 also reported similar results involving 11% of all patients showing increased CRP after POD3, with only 35% of patients carrying increased CRP manifesting SSI, and the origin of inflammation was unknown in 34% of these patients. Therefore, elevated CRP at POD7 and POD14 can give us the possibility to suspect SSI, but we cannot decide SSI. The change rate of CRP on POD14 informed us a reliable validity of detecting SSI. Sensitivity is 83.3%, specificity is 82.4%, respectively.

Based on the normal value (5 mg/L) of CRP, the increase in CRP without specific origin of inflammation frequently occurs. Postoperative elevation of CRP frequently occurred. Elevated CRP was detected in 100 patients on POD7 and 81 patients on POD14. In our study, the number of surgical level varied from 1 to 4. In extensive surgeries, which release a large amount of CRP, the degradation of high plasma levels of CRP to the normal range occurs over a longer period of time. It seems logical that the variation in CRP level instead of simple quantification was used to detect SSI. An elevation in CRP itself was not pathognomonic for SSI. The presence of a secondary infection and other inflammatory processes may trigger a rise in CRP levels. Therefore, it is misleading to diagnose SSI simply according to the increased CRP. When increased CRP value is detected, it is appropriate to exclude other factors that may cause inflammatory reaction. Moreover, the patient’s signs and symptoms, especially pain regardless of the change in posture, should be closely monitored.

SSI was also found in the DCRP group. Even though the value of CRP is decreased on POD7, the suspicion of SSI should not be excluded. There is always a possibility of secondary spike or failure to decrease CRP during the follow-up. One patient even developed SSI without secondary spike or failure to decrease CRP. Twenty-six patients in the DCRP group showed a secondary elevation on POD14, but just 10 patients observed SSI or secondary infection. Sixteen patients were not found any cause for infection. The mean change rate of CRP between POD7 and POD14 cannot show a reliable validity, sensitivity is 90.9% and specificity is 68.1%.

The cut-off values of CRP in the ICRP and the DCRP groups showed NPV higher than 95%. The normalization of CRP until POD14 or decline of CRP within normal limits suggests exclusion of SSI. Meyer et al19 also reported that the serial CRP test showed high sensitivity, specificity, and NPV, but a low PPV of 48.4%. Kang et al14 stated that the detection of SSI based on serial CRP measurement had a sensitivity of 83.3%, specificity of 96.8%, PPV of 31.3%, NPV of 99.7% and normal CRP result obviates the need for further examination. There has been no consensus to detect SSI based on CRP level. A normalization of CRP until POD 14 is essential to almost rule out SSI. However, one patient with normal CRP levels presented with SSI, suggesting that CRP value alone is a supplementary diagnostic parameter. Patients’ signs and symptoms such as fever and aggravation of nocturnal pain regardless of posture are carefully investigated. Secondary infections should be excluded by radiologic and laboratory evaluations. Sometimes, the specific pain can be an important stigma for the diagnosis of SSI. Therefore, close monitoring of the patient cannot be overemphasized. MRI for spine is also considered vigorously in patients suspected with SSI.


The present study has several limitations. First, it is a retrospective study with a small sample size, and therefore associated with an inherent bias. Second, the late SSI was not described in this study. The late SSI is known to be caused by different factors unlike acute SSI, such as hematogenous spread or reactivation of subclinical infection. The diagnosis of late SSI based on serial CRP in acute phase is limited.

Finally, our study was conducted only in patients excluding many underlying diseases causing inflammatory reactions such as primary spondylilitis, autoimmune disease, liver disease, malignancy, and chronic renal disease. However, patients with such underlying diseases are common in clinical practice. Therefore, there might be unpredictable complexity to apply our results to clinical circumference.


Postoperative elevation of CRP frequently occurred. The serial change rate of CRP is a reliable method for detecting SSI more than the criteria of normal value. Moreover, the normalization of CRP until POD14 or decrease of CRP rate within normal limit can rule out early SSI. Nevertheless, close monitoring of patient’s signs and symptoms is important to detect early SSI.

Ethical approval

Institutional review board approval was obtained (2021-0790-0001).

Sources of funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author contribution

M.H.K.: collected and analyzed the data of this study, and drafted the manuscript. J.-H.P.: designed this study and processed data analysis statistically, and revised the manuscript. J.T.K.: reviewed the manuscript and corrected errors.

Conflicts of interest disclosure

The authors declare that they have no financial conflict of interest with regard to the content of this report.

Research registration unique identifying number (UIN)



Jong-Hyeok Park.


1. Fang A, Hu SS, Endres N, et al. Risk factors for infection after spinal surgery. Spine 2005;30:1460–5.
2. Weinstein MA, McCabe JP, Cammisa FP Jr. Postoperative spinal wound infection: a review of 2,391 consecutive index procedures. Clin Spine Surg 2000;13:422–6.
3. Bible JE, Biswas D, Devin CJ. Postoperative infections of the spine. Am J Orthop (Belle Mead, NJ) 2011;40:E264–71.
4. Gerometta A, Olaverri JCR, Bitan F. Infections in spinal instrumentation. Int Orthop 2012;36:457–64.
5. Foglar C, Lindsey R. C-reactive protein in orthopedics. Orthopedics 1998;21:687–91.
6. Black S, Kushner I, Samols D. C-reactive protein. J Biol Chem 2004;279:48487–90.
7. Takahashi J, Ebara S, Kamimura M, et al. Early-phase enhanced inflammatory reaction after spinal instrumentation surgery. Spine 2001;26:1698–704.
8. Thelander U, Larsson S. Quantitation of C-reactive protein levels and erythrocyte sedimentation rate after spinal surgery. Spine 1992;17:400–4.
9. White J, Kelly M, Dunsmuir R. C-reactive protein level after total hip and total knee replacement. J Bone Joint Surg Br 1998;80:909–11.
10. Cunningham ME, Girardi F, Papadopoulos EC, et al. Spinal infections in patients with compromised immune systems. Clin Orthop Relat Res (1976-2007) 2006;444:73–82.
11. Kudo D, Miyakoshi N, Hongo M, et al. Relationship between preoperative serum rapid turnover proteins and early-stage surgical wound infection after spine surgery. Eur Spine J 2017;26:3156–61.
12. Mok JM, Pekmezci M, Piper SL, et al. Use of C-reactive protein after spinal surgery: comparison with erythrocyte sedimentation rate as predictor of early postoperative infectious complications. Spine 2008;33:415–21.
13. Neumaier M, Metak G, Scherer MA. C-reactive protein as a parameter of surgical trauma: CRP response after different types of surgery in 349 hip fractures. Acta Orthop 2006;77:788–90.
14. Kang BU, Lee SH, Ahn Y, et al. Surgical site infection in spinal surgery: detection and management based on serial C-reactive protein measurements. J Neurosurg Spine 2010;13:158–64.
15. Olsen MA, Nepple JJ, Riew KD, et al. Risk factors for surgical site infection following orthopaedic spinal operations. J Bone Joint Surg 2008;90:62–9.
16. Fujita R, Takahata M, Kokabu T, et al. Retrospective study to evaluate the clinical significance of a second rise in C-reactive protein level following instrumented spinal fusion surgery. J Orthop Sci 2019;24:963–8.
17. Mok JM, Guillaume TJ, Talu U, et al. Clinical outcome of deep wound infection after instrumented posterior spinal fusion: a matched cohort analysis. Spine 2009;34:578–83.
18. Iwata E, Shigematsu H, Koizumi M, et al. Lymphopenia and elevated blood C-reactive protein levels at four days postoperatively are useful markers for early detection of surgical site infection following posterior lumbar instrumentation surgery. Asian Spine J 2016;10:220.
19. Meyer B, Schaller K, Rohde V, et al. The C-reactive protein for detection of early infections after lumbar microdiscectomy. Acta Neurochir 1995;136:145–50.

Surgical site infection; C-reactive protein; Posterior lumbar surgery; Postoperative infection

Copyright © 2021 The Authors. Published by Wolters Kluwer Health, Inc. on behalf of IJS Publishing Group Ltd.