|Year : 2002 | Volume
| Issue : 10 | Page : 501-507
C-reactive protein (CRP) as an indicator of sepsis in orthopaedic trauma
R Gupta, R Singh, M Soni
Department of Orthopaedic Surgery Pt. B.D. Sharma PGIMS, Rohtak 124001, India
42/9J, Medical Enclave Rohtak 124001, Haryana
Source of Support: None, Conflict of Interest: None
|How to cite this article:|
Gupta R, Singh R, Soni M. C-reactive protein (CRP) as an indicator of sepsis in orthopaedic trauma. Indian J Med Sci 2002;56:501-7
Physical trauma is a leading cause of mortality and morbidity in all age groups, with post-traumatic or postoperative infection not an infrequent complication.  Therefore, a parameter helpful in early diagnosis of infection, is not only advantageous for early institution of treatment but also prevents morbidity to a large extent.
Ellitsgaard et al observed CRP to parallel the clinical course of inflammatory and infectious diseases and tissue necrosis.  Other tests reported in the literature as indicators of sepsis include ESR, WBC count, body temperature and pulse rate.  However, Yoon et al have reported CRP to be a more sensitive and consistent indicator than ESR for early diagnosis of infection and its response to treatment.  CRP was found to be an early and reliable indicator of sepsis in replacement arthroplasties. ,,,, However, not many reports are available to indicate the usefulness of CRP as an early indicator of sepsis in surgical procedures following orthopedic trauma. ,, It is more so because trauma, surgical or otherwise, itself produces a rise in CRP, which may be superimposed by rise of CRP on account of subsequent sepsis.
The present study was therefore conducted to define the degree and timing of natural response of CRP associated with soft tissue or bony trauma and to find whether CRP is a good indicator of sepsis in surgery following orthopaedic trauma.
| ¤ Material and Methods|| |
The study was conducted in 100 patients of all age groups having a variety of orthopaedic trauma as under Group A- 25 patients undergoing elective orthopaedic surgery; Group B-25 patients of simple or compound long bone fractures; Group C-25 patients with only soft tissue injuries. All the patients of group B and C reported on the day of trauma itself and were not subjected to any major surgical procedure during the period of study. Group D-25 patients with acute orthopaedic infections. In addition a control group E, consisting of 25 randomly selected normal healthy adult blood donors, was included to take into account the baseline value of various parameters prevalent in the population.
All the patients were subjected to a detailed history, thorough physical examination and investigations to rule out any pre-existing systemic illness. Prophylactic broad spectrum antibiotic was given 45 minutes before surgery to all patients of Group A (elective surgery). All the patients of Group D (infection group) were subjected to drainage of pus followed by appropriate broad spectrum antibiotic therapy.
C-reactive protein was measured preoperatively (Group A) or on first day (Group B, C and D) and subsequently on 3rd, 5th and 7th day. Simultaneous, ESR, TLC and temperature monitoring was also done at every stage. Periodic local examination of the injured part including wound or stitch line, if any, was also performed to assess clinical evidence of infection. In addition patients were subjected to daily check-up to rule out any other possible area of infection like chest infection, thrombophlebilis and urinary tract infection etc.
Blood samples were taken in the control group for a one time measurement of CRP, ESR, TLC besides noting down their temperature.
CRP estimation was done semiquantitativeiy using CRP tests kits with reagent sensitivity of 7µg/ml. Serial dilutions of patient's sera were prepared with physiological saline (0.9%) upto a maximum dilution of 1:64. The last dilution of serum with visible agglutination was considered to be the CRP titre of the serum. The CRP concentration was then calculated by multiplying dilution titre with regent sensitivity.
Data so obtained was analysed and correlated with the clinical and haematological findings to draw appropriate conclusions. Students 't' test was employed to find out the statistical significance.
| ¤ Observations|| |
Age ranged from 6 to 65 years with a marked male preponderance, though the average age in various groups was comparable. None of the parameter observed i.e. CRP, TLC, ESR or temperature showed any correlation with age (r <0.6, p>0.05).
The mean CRP on day 1 (preoperative) in the elective surgery group (Group A; 3.36 µg/ml) was not significantly different from that of control group (2.52µg/ ml; p >0.05). However, mean preoperative value of TLC and ESR in group A was significantly different from the control group (p <0.05). Mean of various parameters (CRP 384µg/ml, ESR 69.12 mm 1st hour, TLC 10308 per c.cm and temperature 101.4°F) in the infective group (Group D) on day 1 were significantly different from control group means (CRP 2.52 µg/ml, ESR 17.18 mm 1st hour, TLC 7724 per c.cm. and temperature 98.4°F; p<0.001). Peak values of CRP observed in different groups were, group A 340.5±127.2 µg/ml; group B 376.3±106.6µg/ml; group C 131±104.9 µg/ml and group D 384±136 µg/ml.
On analyzing the profiles of various parameters following tissue trauma (Surgical or accidental) in groups A, B and C, it was observed that the CRP level increased sharply upto day 3. Similarly ESR, TLC and temperature also showed a rise from day 1 to day 3. From day 3 to day 5, while CRP declined sharply in all the three groups, decline in TLC and temperature was not marked. On the other hand, ESR showed a decline in elective surgery and soft tissue injury groups, but showed a rise in fracture group. From day 5 to day 7, CRP further declined in all the groups. Similarly, ESR, TLC and temperature also showed a slow declined in all the groups. However, the profile of various parameters in infection group (Group D) having high values on day 1 showed decline of all the parameters subsequently.
In elective surgery group, the mean duration of surgery was 78.4±30 minutes with range of 25 to 180 minutes. A statistically significant positive correlation (r >0.6, p <0.05) was observed between the duration of surgery and the CRP levels on day 3,5 and 7 after surgery [Table 1]. TLC and ESR on the other hand were found to have no correlation with the duration of surgery (p >0.05). Temperature had a limited correlation only in early postoperative period (day 1-3).
A total of 10 patients out of 75 (Group A, B and C) suffered from septic complication in the study period. The mean CRP in non-infected patients showed a rise of 35.67%, while infected patients showed a rise of 4967% from day 1 to day 3. From day 3 to day 5, the non-infected patients showed a decline of 71 %, whereas decline in infected patients was only 26%. From day 5 to day 7, the non infected patients showed a further decline of 51.4%, while infected patients showed a secondary rise of 14.29%. Profile of other parameters did not show any significant difference. However, ESR showed a peak of 76.2±26.69 mm 1st hour on day 5, TLC 8600±1606 per c.cm. on day 5 and temperature 100.6±0.84°F on day 3. The patients who showed a persistent rise in CRP level or showed a secondary rise in CRP level after day 3 were considered CRP positive and who showed continuous decline in CRP level after day 3 were considered CRP negative. Seven out of the10 infected patient of group A, B and C (n=75), were CRP positive (true positive) and 3 were CRP negative (false negative). Out of the remaining 65 patients, 62 patients were true negatives and 3 false positive. Positive predictive value calculated was 70% and negative predictive value 95.3%.
Subsequent to the management of infection group (Group B) in the form of drainage procedure and appropriate medications, 12 patients showed clinical improvement by day 7, with no drainage of pus (improved). Remaining 13 patients continued to discharge pus even on day 7 inspite of treatment (non-improved).The difference in the mean CRP values of these two subgroups was observed to be significant on day 1 (p <0.05) and highly significant on all subsequent days (p <0.001). Similarly, the difference of mean TLC and temperature was also significant in the two subgroups (p value .01 to .001). On the other hand the difference in the mean ESR of two groups was observed to be significantly only on 7th day.
Average amount of pus drained from the patients (Group D) was observed to be 97.8 ml (10-750ml). Amount of pus drained on any of the day did not have any statistically significant correlation with the different parameters observed.
| ¤ Discussion|| |
C-reactive protein is an acute phase protein and has been considered a non specific indicator of tissue insult. It is, therefore, important to know the natural response of CRP after accidental or surgical trauma before its value can be used to diagnose septic complication in procedures following orthopaedic trauma.
The various groups were made in the present study, to account for different types of tissue trauma. The CRP levels were measured using semi-quantitative latex agglutination method because it is rapid, less expensive and readily available method, as compared to other quantitative methods. Simultaneous, ESR, TLC and temperature monitoring was also done to assess their natural response to the tissue trauma and as to whether they can also be used as indicators for subsequent septic complication. Additional group of active orthopaedic infections was taken to assess as to whether peak values of indicators are different in infective and non-infective tissue insult.
No correlation with age was found with any of the parameters observed in the present study. Similar findings have also been observed by Larsson et al. 12 Absence of any statistically significant difference between mean CRP of group A (preoperative) and group E indicates absence of any pre-existing cause affecting CRP However, mean preoperative values of TLC and ESR in the elective surgery group were significantly different from control group (p <0.05), indicating wide variability of their values even in the normal population. Sanzen and Carlsson, in their study, have also reported that normal range of ESR is wide and values from healthy and infected patients can overlap, thus leading to difficulties in interpretation .  Aalto et al and Shih et al also have observed similar findings in their reports. ,
A significant positive correlation between CRP and duration of surgery on all the subsequent days (p <0.05) in group A indicates that CRP is a very sensitive indicator of tissue damage, irrespective of its nature. These findings are further reinforced by the fact that average peak CRP value attained in the long bone fractures (376.3 µg/ml, was much higher as compared to the soft tissue injury group (131 µg/ml). Similar average peak CRP values after different types of tissue trauma viz. surgery (340.5±127.2,µ g/ml), fractures (376.3±106.6µg/ml) and infection group (384±136 µg./ml) also indicates that CRP does not differentiate between various types of tissue trauma and hence is a non specific indicator. Dominioni et al have also reported a close correlation between the rise in CRP level and the magnitude of the operation .  Kallio et al have also stated that open reduction and plating resulted in a greater CRP response than closed intramedullary nailing in tibial fractures indicating additional tissue trauma caused by open surgery.  The present study however does not corroborate findings of Larsson et al indicating no significant correlation between duration of surgery and CRP levels.  Very high post-traumatic levels of CRP as an indicator of extensive tissue insult can be used to initiate appropriate measures, as extensive tissue damage increases the risk of post-traumatic infections.
Consistent marked rise of CRP levels from day 1-3 followed by its equally sharp and consistent decline confirms it to be a very sensitive indicator of tissue trauma with a short half life. In contrast, rise of ESR, TLC and temperature was observed to be slow, reaching its peak on variable days (day 3-5) followed by a slow decline, indicating their poor sensitivity in respect of tissue trauma. Similar profile of CRP has also been documented by Fisher et al and Shih et al . , Aalto et al have also reported that CRP show a rapid postoperative rise reaching maximum level on 2nd postoperative day followed by a sharp decline. They further observed that ESR values reached their peak value at variable period upto the end of 1st postoperative week followed by slow decline.  This aspect of the CRP profile can be used to diagnose postoperative and post-traumatic septic complication. Subsequent rise in the CRP levels or persistently high values beyond third day should alert the surgeon of possible septic complication. The same cannot be said with certainty about other parameters (TLC, ESR and temperature) because of their smaller rise, peak on variable days and subsequent smaller decline which might conceal a possible infectious complication. Moreover wide variation of TLC and ESR in normal population makes it difficult to interpret their absolute postoperative values.
All the patients who developed septic complication in groups A, B or C showed a significant high CRP level on day 3 as compared to rest of the patients indicating that these patients had suffered a more extensive tissue insult. While patients not complicated by infection in these groups showed a consistent sharp declined after day 3, the patients who developed septic complications showed a s0sequent rise after day 5.
Therefore, the present study confirms the findings of reports in the literature that CRP was superior to ESR, WEB count and the diagnosis based on clinical signs, for early diagnosis of postoperative infection . , Niskanen et al have also reported that a pronounced rise in CRP level on the second and third postoperative day after a major orthopaedic operation is normal but a further rise at one or two weeks suggest the presence of a serious complication . 
Present study confirms the findings if Mustard et al that CRP has a high negative predictive value in ruling out postoperative infection.  Higher negative predictive value of CRP indicates the sharp decline of CRP level beyond third day after trauma not followed by any subsequent rise does ensure the exclusion of any post-traumatic septic complication.
With the initiation of treatment in the infection group, all the parameters showed decline in their values, it being more marked in patients who showed improvement with treatment the decline of CRP was however more marked and well defined as compared to other parameters. Pettola et al have also reported that the normalisation of CRP levels in patients with septic arthritis suggests true extinction of actively invasive bacterial process and ESR is of little use in estimating the prognosis of septic arthritis.  No correlation was found between the amount of pus drained and any of the parameters it can possibly be explained by the fact that a large amount of pus localized to an area may not cause as much tissue damage, as compared to a widespread infection causing extensive tissue damage even if the total amount of pus drained is less.
CRP therefore appears to be a very sensitive though nonspecific indicator of tissue trauma. However, its consistent sharp decline by 3rd day following aseptic tissue trauma can be used with confidence to predict absence of any subsequent septic complication. Moreover, it can also be used as an useful indicator of efficacy of treatment following septic complication. The same however cannot be said about the other commonly used parameters on account of either their wide variability of normal values (TLC & ESR) or less marked rise, variable peak (3-5 days) followed by a slow decline even following aseptic tissue trauma.
| ¤ Summary|| |
In the present study serial estimations of changes in plasma C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), total leucocyte count (TLC) and temperature were recorded in 100 patients. Nature and extent of tissue trauma varied from fresh soft tissue or bony injuries to elective orthopaedic procedures, besides acute orthopaedic infections. All the parameters showed a rise from day 1-3 from the normal levels, though the rise in temperature was minimal. However, from third day onwards, CRP level showed a sharp decline in all cases without any septic complication or infective patients who improved with treatment. Decline of ESR levels on the other hand was variable. Similarly, decline of TLC and temperature was also not consistent and sharp. Persistent rise of CRP level beyond third day or any subsequent rise in CRP level was consistent with a septic complication in the patient. CRP was therefore observed to be a sensitive and dependable indicator of orthopaedic sepsis even orthopaedic trauma.
| ¤ References|| |
|1.||el Hassan BS, Peak JD, Wicher JT, Shepherd J. Acute phase protein levels as an index of severity of physical injury. !nt J oral Maxillofac Surg 1990; 19: 346-349. |
|2.||Ellitsgaard A, Andersson AP, Jensen KV, Jargenson M. Changes in C-reactive protein and erythrocyte sedimentation rate after hip fractures. lnt Orthop 1991; 15: 311-314. |
|3.||Kallio P, Michelesson JE, Lalla AM, Holm T. C-reactive protein in tibial fractures. Natural response to injury and operative treatment. J Bone Joint Surg 1990; 72B: 615-617. |
|4.||Yoon SI, Lim SS, Rha JD et al. C-reactive protein in patients with long bone fractures and after arthroplasty. lnt Orthop 1993; 17: 198-201. |
|5.||Aalto K, Osterman K, Peltola H, Rasanen J. Changes in erythrocyte sedimentation rate and C-reactive protein after total hip arthroplasty, Clin Oil hop 1984; 184:118-120. |
|6.||Niskanen RO, Korkalo 0, Pammo H. Serum C-reactive protein levels after total hip and knee arthroplasty. J Bone Joint Surg 1996, - 788:431-433. |
|7.||Sanzen L, Carlsson AS. The diagnostic value of C-reactive protein in infected total arthroplasties. J Bone Joint Surg 1989; 718: 638-641. |
|8.||Shlh L-Y, Wu J-J, Yang D-J. Erythrocyte sedimentation rate and C-reactive protein values in patients with total hip arthroplasty. Clin Orthop 1987, , 225. , 238-246. |
|9.||Waleczek H, Kozianka J, Everts H. C-reactive protein in the early diagnosis of postoperative infection following bone surgery. Chirurg 1991; 62(12):866-870. |
|10.||Ernest JM, Swain M, Block SM et al. C-reactive protein: A limited test for managing patients with preterm labour or premature rupture of membranes. Am J Obstet Gynecol 1987; 156:449-454. |
|11.||Phillip AGS, Andrews PA. Rapid determination of C-reactive protein levels. Semiquantitative versus quantitative. J Paediatr 1987; 110: 263-265. |
|12.||Larsson S,Thelander U, Friberg S. C-reactive protein levels after elective orthopaedic surgery. Clin Orthop 1992, , 275. - 237-241. |
|13.||Dominioni L, Dionigi R, Cividini F. Determination of C-reactive protein and alpha 1-antitrypsin for quantitative assessment of surgical trauma. EurJ Surg Res 1980; 1(Suppl): 133-135. |
|14.||Fischer CL, Gill C, Forrester MG, Nakamura R. Quantitation of "Acute phase proteins" postoperatively: value in detection and monitoring of complications. Am J Clin path 1976; 66:840-846. [PUBMED] |
|15.||Mustard RA Jr, Bohnen JMA, Haseeb S, Kasina R. C-reactive protein levels predict postoperative septic complications. Arch Surg 1987; 122:69-73. |
|16.||Peltola H, Vahvanen V, Aalto K Fever, C-reactive protein and erythrocyte sedimentation rate in monitoring recovery from septic arthritis: a preliminary study J Pediatr Orthop 1984;4:170-174. |
|This article has been cited by|
||Similar erythrocyte sedimentation rate and C-reactive protein sensitivities at the onset of septic arthritis, osteomyelitis, acute rheumatic fever
| ||Reitzenstein, J.E., Yamamoto, L.G., Mavoori, H. |
| ||Pediatric Reports. 2010; 2(1): 32-35 |
||Diagnostic accuracy of C-reactive protein for intraabdominal infections after colorectal resections
| ||Křrner, H., Nielsen, H.J., Sřreide, J.A., Nedrebř, B.S., Sřreide, K., Knapp, J.C. |
| ||Journal of Gastrointestinal Surgery. 2009; 13(9): 1599-1606 |
||Diagnosing wound infection: The use of C-reactive protein
| ||Kingsley, A., Jones, V. |
| ||Wounds UK. 2008; 4(4): 32-46 |