Introduction

The mortality rate of colorectal surgery is almost 35% and a significant proportion corresponds to intraabdominal complications¹ which usually occur between postoperative days 5 and 7, when oral feeding has started and, occasionally, the patient has been discharged². In the context of moving towards laparoscopic minimally invasive surgeries and ERAS or “fast-track” protocols, length of hospital stay has been reduced and patients are discharged from hospital earlier³. Several studies have focused on finding early indicators of complications to identify patients who deviate from the expected course and, thus, individualize the conditions upon discharge⁴.

C-reactive protein (CRP) is part of the systemic inflammatory response and has a relatively short half-life, so its levels can vary in 24 to 48 hours. The presence of CRP indicates an acute inflammatory reaction and plays a fundamental role in maintaining acute inflammation⁵. The possibility of measuring CRP in blood and its rapid response to inflammation make PCR an attractive marker for early detection of complications.

The aim of our study was to investigate the performance of CRP in the postoperative period following scheduled colorectal surgeries and its association with the early diagnosis of complications.

Material and methods

All the patients consecutively undergoing major scheduled colorectal surgery between January 2019 and March 2020 were included. Major colorectal surgery was defined as colonic or rectal resection or a combination of both, with or without anastomosis. C-reactive protein was routinely measured starting on postoperative day 3 and recorded in a prospective database. Emergency surgeries, ostomy preparation as a single procedure, surgery for anal orifice or transanal surgery, surgeries due to inflammatory bowel disease or suspected preoperative infection were excluded.

The study design was observational and retrospective.

The personal data, medical and surgical history, postoperative complications, and length of hospital stay were recorded. Those cases of laparoscopic surgery converted to open surgery were considered in the conventional surgery group.

Results

During the study period 2205 surgeries were performed: 938 corresponded to urgent procedures other than colorectal surgeries. Of the remaining 1367 surgeries, 253 were colorectal procedures and 99 of these were major scheduled colorectal surgeries. Thirty patients were excluded due to lack of complete CRP records for the analysis (Fig. 1).

Figure 1 Flow diagram of patient selection 

Mean age was 59 years (SD 13.6, range 33- 85) and 43% were men. The surgeries performed are shown in Table 1. Thirty-three were laparoscopic surgeries, 43 were conventional surgeries and 3 were laparoscopic procedures converted to conventional surgery. Mean length of hospital stay was 10.8 days (SD 10, range 5-70); patients without complications in the conventional surgery group were hospitalized for 9.3 days (SD 4.25) and 6.7 days (SD 1.2) in the laparoscopy group.

Tabla 1 

Postoperative outcomes

Of 69 patients, 7 (10.14%) presented major complications (Table 2). All the patients were operated on for colorectal cancer. Anastomotic leak (fistula, fluid collection) was the most common complication, followed by surgical site complications (hematoma, evisceration). Of the 7 patients, 4 were men with ASA grade II or III. In this subgroup, mean length of hospital stay was 22.3 days (range 8-70). Four patients required reoperation: evacuation of hemoperitoneum, incisional hernia repair with polyglactin mesh, dismantling of anastomosis and enterolysis. Of the remaining three patients, one required surgical drainage of the surgical site, and two underwent percutaneous drainage. There were no deaths in the series.

Table 2 

Group without complications

Mean peak CRP of day 3 was 11.16 mg/dL (range 1.43-98.8) and then decreased until reaching normal values or until the patient was discharged if CRP did not increase between day 5 and 7. On postoperative day 5, mean CPR level was 6.38 mg/dL (range 0.84-17.6) (Fig. 2).

Figure 2 CRP curve in patients without complications from postoperative day 3. See the descent pattern of CRP levels after the peak on day 3. 

In patients undergoing laparoscopy, median CRP levels in the early and late period were 12.87 mg/dL (range 0.84-17.60) and 7.25 mg/dL (range 1.2- 16.9), respectively. In patients undergoing conventional surgery, median CRP levels in the early and late period were 11.43 mg/dL (range 1.43-98.8) and 5.49 mg/ dL (range 2-44.4), respectively, with no significant differences between both groups.

CRP in patients with complications

CRP values were higher than those in patients without complications throughout all the postoperative days (Fig. 3).

Figure 3 CRP curve in patients with and without complications 

Median CRP levels in the early and late period were 18.22 (range 11.90-83.40) and 28 (range 20- 45.9), respectively, and this difference was statistically significant (day 3, p < 0.05; CI: 2.834-32.68; day 5 p < 0.001; CI: 11.24-39.61) (Fig. 4). In all the cases with complications, CRP either increased or did not decrease before symptoms and signs developed.

We found significant differences in CRP values on day 5 but not on day 3 in patients who developed complications, independently of the type of approach (laparoscopy or conventional surgery) (Fig. 5).

Figure 4 C-reactive protein (CRP) values in the early group (postoperative day 3) and late group (postoperative day 5) for patients with and without complications Data are expressed in median and range.  

Figure 5 C-reactive protein (CRP) values in the early group (postoperative day 3) and late group (postoperative day 5) for patients with and without complications by type of approach. Data are expressed in median and maximum and minimum.  

The Spearman’s correlation coefficient between white blood cell count and CRP levels, showed a correlation (r) of 0.52 for the early period and of 0.35 for the late period, without a significant difference (Fig. 6).

Figure 6 Spearman’s correlation coefficient between C-reactive protein (CRP) and white blood cell (WBC) count on day 3 and 6 

Likewise, 6 patients with CRP levels > 10 mg/dL on postoperative day 6 did not show complications on discharge, but they all had a downward curve compared with the previous days.

None of the patients with complications had normal values on day 3 (median 18.22 mg/dL, range 11.90-83.4 mg/dL) or on day 5 (median 28 mg/dL, range 20.6-45.9 mg/dL).

In the analysis of the ROC curve, a cut-off value of 10.92 mg/dL had a sensitivity of 87.50% (CI: 71.93-95.03) and a specificity of 100% (CI: 60.97- 100) to exclude complications on postoperative day 5.

Discussion

The synthesis of CRP mainly occurs in hepatocytes in response to increased levels of interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α) and IL-1-b⁷. CRP binds to immunoglobulin gamma receptor called Fc of leukocytes and to endothelial cells via the activation of IL-88. When the inflammatory process begins, CRP plasma concentration increases during the first 6 hours and reaches a maximum peak between 24-72 hours, but rapidly decreases as inflammation resolves⁸. Its short half-life (19 hours) allows for early monitoring and confers advantages over other acute-phase proteins, such as albumin or ferritin which has longer half-life. Plasma levels of CRP are independent of diet, circadian rhythm or medications⁹. Adamina et al. reported that CRP levels measured on any postoperative day after major abdominal surgeries in patients without postoperative infections were lower and decreased more rapidly from postoperative day 3 than in patients with postoperative infections¹⁰. Peak levels of CRP were reached on postoperative day 2 for patients without complications and on postoperative day 3 for patients with complications and declined on day 5.

Sustained high levels are an indicator of postoperative complications and their decline has high negative predictive value for the development of complications at discharge¹. A particular kinetics known as “increase and decrease pattern” is identified in patients without complications and “a steady rise or second rise” is seen in those with complications¹¹. In our series, the CRP curve in the postoperative period without complications descended on postoperative day 3 until reaching normal values, while patients with complications had CRP levels above normal values (progressive increase or lack of decrease) during the postoperative course until complications resolved.

Colorectal surgery has a morbidity rate of 30% and a mortality rate of 3 to 4%¹². Complications occur from the immediate postoperative period until postoperative day 15 or beyond but are most common between postoperative day 3 and 10¹³. The implementation of accelerated recovery protocols and minimally invasive surgeries has significantly reduced length of hospital stay. The development of a complication after the patient has been discharged involves readmission and, probably, treatment delays.

Increased CRP values or lack of return to normal levels after colorectal anastomosis have been related with the development of anastomotic leakages¹⁴. CRP levels < 140 mg/L on the postoperative day 3 excludes the development of fistula with a sensitivity of 78% and a specificity of 86%¹⁵. Welsh et al. reported that CRP levels of 140 mg/L on postoperative day 3 are useful to differentiate patients with complications from those without complications¹⁶. Kørner et al. established that a cut-off value of 190 mg/dL for CRP on day 3 was associated with intraabdominal complications and identified 4 of 5 patients with fistulas¹⁷. In their study, Pantel et al. determined a cut-off value of 127 mg/dL on postoperative day 3 for patients without complications, of 145mg/L for readmissions and of 147 mg/dL for the development of fistula, with a negative predictive value of 93% and 99%, respectively¹⁸. Benoit et al. described that CRP value <100 mg/dL on postoperative day 3 had a sensitivity of 85% and specificity of 33% to predict the development of complications; on day 5, sensitivity was 72% and specificity was 75%, with an area under the ROC curve of 0.81, a negative predictive value (NPV) of 89.4% and positive predictive value (PPV) of 49%. In the meta-analysis by Warschkow et al., the specificity of CRP to detect complications increases throughout the postoperative period, with the highest sensitivity on day 419.

In our series, we found significant differences in CRP levels in patients with and without complications in the global analysis of the sample. When the analysis was performed according to the approach used, we only found a significant difference on CRP measured on day 5, both for patients operated by laparoscopy and conventional surgery. These results may be due to the wide dispersion of CRP values on day 3, to the impact of surgical inflammation even on day 3, or to the fact that many of the complications develop later in the postoperative period. We established a cut-off value of CRP of 10.92 mg/dL on postoperative day 5 to rule out complications with a sensitivity of 87.5% and specificity of 100%.

As CRP levels increase before the clinical presentation of the complication, increased levels may indicate the need for tests as CT scan for early detection during postoperative days 3 to 5. This approach has been proposed by Straatman⁴ and Adamina, who reported that CRP levels < 5.6 mg/dL on day 4 have a NPV of 100% for the development of complications and allow early discharge. If this value is > 12 mg/dL, a computed tomography (CT) scan is performed even in the absence of clinical signs of infection and, if the value is between these two limits, the patient remains hospitalized and monitored²⁰.

When we compared both approaches, we did not find significant differences and we assume that this is a consequence of the sample size. Other groups, as Pedrazzini et al. reported that CRP levels on postoperative day 3 were almost twice as high in patients undergoing conventional surgery versus those undergoing laparoscopic surgery (205 vs. 104 mg/L), and that a cut-off value of 120 mg/L excluded complications, with a NPV of 97.7%²¹. Straatman et al. also reported lower CRP levels after laparoscopic surgery¹.

Likewise, multiple studies have analyzed its value as a negative predictor for discharge¹⁹. The ascent-descent curve associated with normal values between postoperative days 5 and 7 is highly safe. However, the opposite is not valid, since the positive predictive value is between 27 and 50%, depending on the day on which it is measured¹⁹. Some authors decide hospital discharge based on the absolute CRP value, while others advocate CRP variations. Thus, the absence of a decline or a second rise, even within normal values, are signs of alarm. In our report none of the patients with complications had CRP < 14 mg/dL on postoperative day 5, and those without complications who had abnormal CRP on day 5 had a descent curve in serial CRP determinations.

Finally, in agreement with Benoit et al., we did not find any association between white blood cell count and the development of complications, or between CRP values and leukocytes¹¹. Almeida et al. found that white blood cell count was different between patients with and without fistulas only on day 7, when most fistulas were already clinically evident, compared with early elevation of CRP in patients with fistulas¹⁵. In this sense, the use of CRP as a negative predictor of complications is better than white blood cell count.

Weaknesses and limitations of the study

The retrospective nature of the analysis, the small number of patients with complications, and the fact that the attending surgeons were not blinded to the CRP values, are some limitations of our report.

The exclusion of patients with incomplete records of CRP in the postoperative period could promote selection bias in the cases studied.

Conclusion

CRP values in the postoperative period of major colorectal surgery increase until the third day and then decrease to normal levels in patients without complications. In patients with complications operated on either by laparoscopy or conventional surgery, CRP values on day 3 and 5 are higher than those of patients without complications. Thus, CRP can be used as a marker or predictor of postoperative complications. We believe that routine determination of CRP can help in early detection of complications and provide safe discharge of patients.