- Email: [email protected]
Pyrexia following total knee replacement
The Knee 13 (2006) 324 – 327 www.elsevier.com/locate/knee
Pyrexia following total knee replacement Subhajit Ghosh, Richard M. Charity ⁎, Saadallah G. Haidar, Binod K. Singh Department of Trauma and Orthopaedic Surgery, City Hospital, Dudley Road, Birmingham B18 7QH, UK Received 7 August 2005; received in revised form 1 May 2006; accepted 3 May 2006
Abstract This study aims to determine the incidence and factors associated with pyrexia following total knee replacement (TKR). We performed a retrospective analysis of the temperature charts and histories of patients who underwent 170 TKRs. There was a statistically significant increase in mean temperature from pre-operation to post-operation, and this increase remained significant through to 5 days post surgery (p < 0.0001). Sixty-two (36.5%) patients were pyrexial (≥ 38 °C) at some point. Fourteen patients developed a clinical infection, but only four of these were pyrexial. There was no association between pyrexia and infection, allogenic blood transfusion, haemoglobin loss, use of urinary catheter, rheumatoid arthritis, anaesthetic type, and previous pyrexia following TKR. Pyrexia as a diagnostic test for the development of infection had a sensitivity of 0.286 (95% CI = 0.084–0.581), specificity of 0.628 (95% CI = 0.548–0.704) and positive predictive value of 0.065 (95% CI = 0.018–0.157). Pyrexia in the first 5 days following TKR is usually a normal physiological response and should not cause undue concern about the presence of infection. © 2006 Elsevier B.V. All rights reserved. Keywords: Knee; Arthroplasty; Replacement; Pyrexia; Fever
1. Introduction
2. Patients and methods
The development of pyrexia during the first few days following total knee replacement (TKR) is a relatively common finding [1–3]. Pyrexia is a source of concern for both medical staff and patient as it is considered to be an indicator of possible infection. Patients who develop postoperative pyrexia are therefore often subjected to multiple investigations in order to find a cause for the fever. Empirical treatment with antibiotics is sometimes started and discharge from hospital can be delayed. This study sets out to determine the incidence of pyrexia following TKR and to determine which, if any, factors can be associated with the development of pyrexia in the early postoperative period. We aim to determine whether the concern about pyrexia following TKR and its association with infection is justified.
The medical records of the patients undergoing 178 consecutive primary unilateral TKR in our unit were reviewed retrospectively. The records for eight episodes were substantially incomplete or missing and these were excluded from the study, leaving a study group of 170 TKRs. These operations were performed on a total of 121 different individual patients (72 patients had unilateral TKR, while 49 patients had bilateral TKR at separate sittings at least 3 months apart). There were 67 females and 54 males. There were 88 operations on the right knee and 82 on the left knee. Mean age at operation was 68.5 years (range 47 to 88 years). Rheumatoid arthritis was the indication for surgery in 17 TKRs (in 11 different patients); in all other cases the diagnosis was osteoarthritis. All patients were operated on in a single hospital using the PFC Sigma knee system (DePuy Orthopaedics Inc., Warsaw, IN, USA) by any one of five consultant orthopaedic surgeons. All operations were performed in an operating theatre with a laminar airflow system and Charnley exhaust suits were universally used. All surgery was performed under a tourniquet and the implants were cemented using polymethylmethacrylate cement containing antibiotic (Palacos R with gentamycin (Biomet Inc., Warsaw, IN, USA)). Antibiotic prophylaxis of 750 mg Cefuroxime, given intravenously (one dose at induction of anaesthesia plus three post-operative doses), was received by all. Either one or two suction drains were used and, according to surgeon preference, were removed on either the first or post-operative day. DVT prophylaxis (compression stockings plus either low-molecular-weight heparin
⁎ Corresponding author. 25 Lyndhurst Road, Crosby, Liverpool, L23 9TN, UK. Tel.: +44 1519240495. E-mail address: [email protected] (R.M. Charity). 0968-0160/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.knee.2006.05.001
S. Ghosh et al. / The Knee 13 (2006) 324–327
325
Table 1 Summary of all patients who had an infection diagnosed Infection type
Number of patients
Number of patient who had pyrexia during study period
Deep wound (prosthesis) Superficial wound Respiratory tract Urinary tract
2
0
8 3 1
3 1 0
or aspirin, depending on surgeon preference) was given until discharge from hospital. Patients were seen by a physiotherapist and mobilization commenced on the first post-operative day. Patient body temperature was routinely measured, three times daily (at 07:00, 13:00 and 19:00 h), by a trained nurse, using an infra-red tympanic thermometer and recorded on an observation chart which was filed in the patient's medical records. After TKR patients are routinely followed-up in the outpatient clinic for at least 5 years. Data collected from the medical records were: maximum body temperature recorded on each calendar day (pre-operative day, day of operation (day 0) and for 5 days (days 1–5) post-operatively or until discharge), whether an infection was diagnosed (any infection as an inpatient/infection of wound or prosthesis at any time during follow-up), whether a post-operative allogenic blood transfusion was given, preoperative and second post-operative day haemoglobin level, whether a urinary catheter was used, type of anaesthetic (general or regional) used and date of last outpatient follow-up. Data for pre-operative temperature (in 11 cases), day of operation temperature (in 5 cases), anaesthetic type (in one case) and second post-operative day haemoglobin level (in 63 cases) were missing from the medical records. For the purpose of our study, pyrexia is defined as a recorded body temperature of 38 °C or greater. Surgical wound infection, either superficial or deep, was diagnosed by the criteria suggested by the National Center for Infectious Diseases, Centers for Disease Control and Prevention, USA [4]. A urinary tract infection was defined as a positive urine culture that was treated with antibiotics. A respiratory tract infection was defined as either consolidation on chest X-ray and/or positive sputum culture that was treated with antibiotics. Statistical analysis of the association between pyrexia and infection, allogenic blood transfusion, haemoglobin loss, use of urinary catheter, rheumatoid arthritis, anaesthetic type, and previous pyrexia following TKR was by Fisher's Exact Test. Analysis of post-compared to pre-operative temperature was by Wilcoxon matched-pairs signed-ranks test. Statistical significance was set at p < 0.05. Data analysis and calculations were performed using the InStat statistical software package, version 3.0b for Macintosh (GraphPad Software, San Diego, CA, USA).
Fig. 1. Daily maximum temperature following TKR – means with 95% CI (box) and ranges (whiskers) shown.
Two patients developed a prosthetic infection. The first was a 60year-old woman with osteoarthritis who complained of increasing knee pain 6 months post TKR, in whom a Staphylococcus aureus prosthetic infection was diagnosed. She ultimately went on to have removal of the implant and arthrodesis of the knee. The second case was a 68-year-old man with rheumatoid arthritis who developed an inflamed knee 3 years after TKR. This was diagnosed as a deep S. aureus infection and was treated by arthroscopic debridement and long-term suppressive antibiotic therapy. There were no cases of DVT diagnosed in the study group. The maximum daily body temperatures are summarized in Table 2 and Figs. 1 and 2. It can be seen from Fig. 2 that there is a statistically significant increase in mean temperature from preoperation to post-operation, and that this increase remains significant through to 5 days post surgery (p < 0.0001). In 62/170 operations (36.5%) pyrexia was present at some point during the first 5 post-operative days. The daily prevalence of pyrexia is shown in Fig. 3, with the highest prevalence on day 1. The association between the various factors under investigation and the development of post-operative pyrexia are summarized in Table 3. There were no statistically significant associations,
3. Results Mean follow-up time was 19.2 months (range 2–72, interquartile range 6–28). There were 14 post-operative infections (infection rate 8.2%), all in different patients, detailed in Table 1. Table 2 Daily maximum temperature distribution (°C) of patients before and for 5 days after TKR Temperature
Pre-op
Day 0
Day 1
Day 2
Day 3
Day 4
Day 5
<36 36–36.9 37–37.9 38–38.9 ≥39 n % Pyrexial
0 146 13 0 0 159 0
11 109 35 10 0 165 6.1
7 53 79 31 0 170 18.2
3 47 92 27 1 170 16.5
1 71 81 17 0 170 10.0
1 92 64 12 0 169 7.1
0 95 61 3 0 159 1.9
Values indicate number of patients.
Fig. 2. Daily maximum temperature following TKR – means with 95% CI shown.
326
S. Ghosh et al. / The Knee 13 (2006) 324–327
Fig. 3. Daily prevalence of pyrexia ≥38 °C following TKR.
although haemoglobin loss of more than 3 g/dl was nearly significant (p = 0.075, odds ratio 2.167, 95% CI = 0.986–4.76). Of the 14 cases that developed a clinical infection, only four were pyrexial. There were 58 cases that developed pyrexia and had no clinical evidence of infection. Using the presence of pyrexia at any time as a diagnostic test for the development of infection gives a sensitivity of 0.286 (95% CI = 0.084–0.581), specificity of 0.628 (95% CI = 0.548–0.704) and positive predictive value of 0.065 (95% CI = 0.018–0.157).
4. Discussion Pyrexia in the first few days following many types of surgery, including joint replacement surgery, is a relatively common finding [1–3,5]. The physiological response to surgical trauma, disturbance of thermoregulation by anaesthesia, retained haematoma, atelectasis, pneumonia, urinary tract infection, superficial or deep wound infection, deep venous thrombosis (DVT) and response to blood transfusion have been suggested as some of the potential causes of a raised temperature in the TKR patient [1,2,6,7]. Amongst these many causes the overwhelming concern in the mind of the surgeon is that there may be a deep infection of the wound or prosthesis as this is associated with significant morbidity and mortality [8,9]. Due to this fear, TKR patients who develop a post-operative pyrexia are often subjected to a routine panel of investigations (chest X-ray, urine, blood and wound-swab cultures, serial full-blood-counts and inflammatory markers, Doppler ultrasound or venogram to exclude DVT) in an attempt to determine the cause of the fever. In our experience, applied blindly, these tests seldom produce useful diagnostic information. In addition, on development of a postoperative fever, additional antibiotics are often prescribed without firm evidence of infection being present, despite prolonged antibiotic use being potentially harmful [10]. The number of publications looking at pyrexia following TKR is quite limited, with some differing results. Kennedy et al. [1] looked at 90 TKR patients and found that 17%
developed a high pyrexia (≥ 39 °C), most commonly on the second post-operative day. They found significant independent associations between high pyrexia and both blood transfusion and fall in haematocrit. They found no association between pyrexia and the presence of an infective focus and concluded that pyrexia in the first few days following TKR is a normal physiological process. Guinn et al. [2] looked at 95 unilateral and 23 bilateral TKRs. They found a post-operative pyrexia (≥ 38 °C) rate of 66% for unilateral and 74% for bilateral TKRs. They found that patients who experienced a fever were significantly more likely to develop a complication in the immediate postoperative period, but there was no difference in the rate of long-term complications. No association between operative blood loss and pyrexia was found. They concluded that the rate of pyrexia after TKR is high and, based on the natural history of the fever, that atelectasis and DVT were the most likely causes. Shaw and Chung [7] examined 100 patients undergoing TKR plus 100 patients undergoing total hip replacement (THR). They found that the mean maximum temperature was highest on the first post-operative day and that 19% had a fever of ≥ 39 °C at some point. No association between blood loss or transfusion and pyrexia was found. Their conclusion was that post-operative fever after total joint arthroplasty is a normal physiological response and that a full work-up for fever alone is not justified. The policy in our unit is to use a four-dose antibiotic prophylaxis regimen and also to routinely use suction drains for all TKRs. Four doses of antibiotic, compared to three or less, has been shown to be effective in reducing infection rates following total hip replacement (THR) [11]. The routine use of wound drains is more controversial. A meta-analysis [12] has shown that, following TKR or THR, the use of drains was associated with a greater need for blood transfusion, yet made no difference to the rates of wound infection, haematoma, limb swelling, venous thrombosis or hospital-stay.
Table 3 Association between various factors and the development of a post-operative pyrexia following TKR
Infection (n = 170)
No Yes Transfusion (n = 170) No Yes Haemoglobin <3 g/dl loss (n = 107) ≥3 g/dl Urinary catheter (n = 170) No Yes Arthritis type (n = 170) Osteo Rheumatoid Anaesthetic type (n = 169) General Regional Previous pyrexia No after TKR (n = 49) Yes
No pyrexia Pyrexia
p
98 10 64 44 39 25 83 25 99 9 72 35 20 10
0.5786 NS 0.518 NS 0.075 NS 0.4421 NS 0.4268 NS 0.7378 NS 0.3774 NS
(58%) (6%) (38%) (26%) (36%) (23%) (49%) (15%) (58%) (5%) (43%) (21%) (41%) (20%)
58 (34%) 4 (2%) 40 (24%) 22 (13%) 18 (17%) 25 (23%) 51 (30%) 11 (6%) 54 (32%) 8 (5%) 40 (24%) 22 (13%) 10 (20%) 9 (18%)
Numbers of patients (and percentages) are shown. NS = not significant.
S. Ghosh et al. / The Knee 13 (2006) 324–327
The rate of deep infection of the prosthesis in our study group was 2/170 (1.2%) which, albeit with a shorter followup time, is comparable to other series [8,9]. Surprisingly, no case of DVT was diagnosed in our study group. This low rate compared to published series may be due to our universal use of chemical prophylaxis and compression stockings, early mobilization and the fact that only patients with significant clinical symptoms were investigated by venography [13]. Our results suggest that there is an element of postoperative hypothermia affecting the temperature of some patients on the day of surgery (temperature < 36 °C in 11/165 patients). Hypothermia in the first 12 h following TKR has been observed previously and the use of measures such as warming blankets and warmed intravenous infusions during surgery are recommended to minimize the problem [14]. In contrast to the studies of Kennedy et al and Shaw and Chung we found that the highest temperatures occurred on the second rather than first post-operative day and also that temperatures of > 39 °C were uncommon. We found no significant association between the development of a pyrexia and blood transfusion, use of urinary catheter, rheumatoid arthritis, anaesthetic type, and previous pyrexia following TKR. Blood transfusion is much less of a risk factor for the development of pyrexia since the change over to using leukocyte-depleted blood and this may partly explain why temperatures of > 39 °C were seldom seen in our cohort of patients [15]. Haemoglobin loss of more than 3 g/dl almost reached significance (p = 0.075) for an association with pyrexia. Haemoglobin loss, independent of subsequent transfusion, has been shown to be a factor associated with fever following TKR in the past and it has been suggested that the pyrexia is a response to haematoma [1,7]. The trend towards an association in our study, together with our universal use of suction drains, suggests that increased tissue trauma and its association with bleeding, rather than haematoma per se, may be the cause of the fever. Studies have suggested that a febrile reaction following TKR is at least partly a physiological response due to surgical site inflammation. Mehra et al. [16] have shown that, following TKR and in the absence of infection, local skin temperature and C-reactive protein level remain elevated for up to 18 weeks and 6 weeks respectively, indicating that a significant and persistent local inflammatory process is taking place. Andreas et al. [3] have shown that the levels of the endogenous pyrogen interleukin-6 in the serum and drain fluid of TKR patients was higher in those who were pyrexial after surgery, suggesting that this cytokine may be involved in the development of the fever. It is interesting to note that only four of the 14 patients in our study who developed an infection were pyrexial in the post-operative period. This may suggest that a lack of a pyrexia is a risk for infection, as the patient has not mounted an adequate immune response.
327
The results of our study support the hypothesis that most post-TKR fevers are physiological, as we found a high pyrexia rate (36%), a significant increase in mean body temperature throughout the 5-day post-operative period, and no correlation between pyrexia and the diagnosis of an infection. The low sensitivity, specificity and positivepredictive-value of pyrexia as a diagnostic test for the presence of infection shown in our study means that, on its own, the presence of pyrexia has little diagnostic value. In the absence of additional clinical findings, the presence of pyrexia in the first 5 days following TKR is usually a normal physiological finding and should not cause undue concern about the presence of infection.
References [1] Kennedy JG, Rodgers WB, Zurakowski D, Sullivan R, Griffin D, Beardsley W. et al. Pyrexia after total knee replacement, a cause for concern? Am J Orthop. 1997(Aug);26(8):549–52, 554. [2] Guinn S, Castro Jr FP, Garcia R, Barrack RL. Fever following total knee arthroplasty. Am J Knee Surg 1999 (Summer);12(3):161–4. [3] Andres BM, Taub DD, Gurkan I, Wenz JF. Postoperative fever after total knee arthroplasty: the role of cytokines. Clin Orthop Relat Res 2003 (Oct);415:221–31. [4] Guidline for prevention of surgical site infection. US Department of Health and Human services, 1999. www.cdc.gov/ncidod/dhqp/pdf/ guidelines/SSI.pdf (accessed March 2006). [5] Than P, Malovics I. Significance of postoperative fever after hip prosthesis implantation. Z Orthop Ihre Grenzgeb 2000 (Sep–Oct);138 (5):430–5. [6] Aburahma AF, Saiedy S. Deep vein thrombosis as probable cause of fever of unknown origin. W V Med J 1997 (Jan–Feb);93(1):368–70. [7] Shaw JA, Chung R. Febrile response after knee and hip arthroplasty. Clin Orthop Relat Res 1999 (Oct);367:181–9. [8] Lentino JR. Prosthetic joint infections: bane of orthopedists, challenge for infectious disease specialists. Clin Infect Dis 2003 (May 1);36 (9):1157–61. [9] Blom AW, Brown J, Taylor AH, Pattison G, Whitehouse S, Bannister GC. Infection after total knee arthroplasty. J Bone Joint Surg Br 2004 (Jul);86(5):688–91. [10] Hanssen AD, Osmon DR. The use of prophylactic antimicrobial agents during and after hip arthroplasty. Clin Orthop Relat Res 1999 (Dec); 369:124–38. [11] Engesaeter LB, Lie SA, Espehaug B, Furnes O, Vollset SE, Havelin LI. Antibiotic prophylaxis in total hip arthroplasty. Acta Orthop Scand 2003 (Dec);74(6):644–51. [12] Parker MJ, Roberts CP, Hay D. Closed suction drainage for hip and knee arthroplasty – a meta-analysis. J Bone Joint Surg Am 2004 (Jun);86-A(6):1146–52. [13] Warwick DJ, Whitehouse S. Symptomatic venous thromboembolism after total knee replacement. J Bone Joint Surg Br 1997 (Sep);79 (5):780–6. [14] Dalen T, Nilsson KG, Engstrom KG. Fever and autologous blood retransfusion after total knee arthroplasty: a prospective study of 40 autotransfusion events in 21 patients. Acta Orthop Scand 2002 (Jun); 73(3):321–5. [15] Zhao SM, Cheng XL, Hu J, Xiang GC, Zhang JS, Li RQ. Clinical assessment of preventing febrile nonhemolytic transfusion reaction by leukocyte-depleted blood transfusion. Zhongguo Shi Yan Xue Ye Xue Za Zhi 2002 (Dec);10(6):568–70. [16] Mehra A, Langkamer VG, Day A, Harris S, Spencer RF. C reactive protein and skin temperature post total knee replacement. Knee 2005 (August);12(4):297–300.
Pyrexia following total knee replacement Subhajit Ghosh, Richard M. Charity ⁎, Saadallah G. Haidar, Binod K. Singh Department of Trauma and Orthopaedic Surgery, City Hospital, Dudley Road, Birmingham B18 7QH, UK Received 7 August 2005; received in revised form 1 May 2006; accepted 3 May 2006
Abstract This study aims to determine the incidence and factors associated with pyrexia following total knee replacement (TKR). We performed a retrospective analysis of the temperature charts and histories of patients who underwent 170 TKRs. There was a statistically significant increase in mean temperature from pre-operation to post-operation, and this increase remained significant through to 5 days post surgery (p < 0.0001). Sixty-two (36.5%) patients were pyrexial (≥ 38 °C) at some point. Fourteen patients developed a clinical infection, but only four of these were pyrexial. There was no association between pyrexia and infection, allogenic blood transfusion, haemoglobin loss, use of urinary catheter, rheumatoid arthritis, anaesthetic type, and previous pyrexia following TKR. Pyrexia as a diagnostic test for the development of infection had a sensitivity of 0.286 (95% CI = 0.084–0.581), specificity of 0.628 (95% CI = 0.548–0.704) and positive predictive value of 0.065 (95% CI = 0.018–0.157). Pyrexia in the first 5 days following TKR is usually a normal physiological response and should not cause undue concern about the presence of infection. © 2006 Elsevier B.V. All rights reserved. Keywords: Knee; Arthroplasty; Replacement; Pyrexia; Fever
1. Introduction
2. Patients and methods
The development of pyrexia during the first few days following total knee replacement (TKR) is a relatively common finding [1–3]. Pyrexia is a source of concern for both medical staff and patient as it is considered to be an indicator of possible infection. Patients who develop postoperative pyrexia are therefore often subjected to multiple investigations in order to find a cause for the fever. Empirical treatment with antibiotics is sometimes started and discharge from hospital can be delayed. This study sets out to determine the incidence of pyrexia following TKR and to determine which, if any, factors can be associated with the development of pyrexia in the early postoperative period. We aim to determine whether the concern about pyrexia following TKR and its association with infection is justified.
The medical records of the patients undergoing 178 consecutive primary unilateral TKR in our unit were reviewed retrospectively. The records for eight episodes were substantially incomplete or missing and these were excluded from the study, leaving a study group of 170 TKRs. These operations were performed on a total of 121 different individual patients (72 patients had unilateral TKR, while 49 patients had bilateral TKR at separate sittings at least 3 months apart). There were 67 females and 54 males. There were 88 operations on the right knee and 82 on the left knee. Mean age at operation was 68.5 years (range 47 to 88 years). Rheumatoid arthritis was the indication for surgery in 17 TKRs (in 11 different patients); in all other cases the diagnosis was osteoarthritis. All patients were operated on in a single hospital using the PFC Sigma knee system (DePuy Orthopaedics Inc., Warsaw, IN, USA) by any one of five consultant orthopaedic surgeons. All operations were performed in an operating theatre with a laminar airflow system and Charnley exhaust suits were universally used. All surgery was performed under a tourniquet and the implants were cemented using polymethylmethacrylate cement containing antibiotic (Palacos R with gentamycin (Biomet Inc., Warsaw, IN, USA)). Antibiotic prophylaxis of 750 mg Cefuroxime, given intravenously (one dose at induction of anaesthesia plus three post-operative doses), was received by all. Either one or two suction drains were used and, according to surgeon preference, were removed on either the first or post-operative day. DVT prophylaxis (compression stockings plus either low-molecular-weight heparin
⁎ Corresponding author. 25 Lyndhurst Road, Crosby, Liverpool, L23 9TN, UK. Tel.: +44 1519240495. E-mail address: [email protected] (R.M. Charity). 0968-0160/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.knee.2006.05.001
S. Ghosh et al. / The Knee 13 (2006) 324–327
325
Table 1 Summary of all patients who had an infection diagnosed Infection type
Number of patients
Number of patient who had pyrexia during study period
Deep wound (prosthesis) Superficial wound Respiratory tract Urinary tract
2
0
8 3 1
3 1 0
or aspirin, depending on surgeon preference) was given until discharge from hospital. Patients were seen by a physiotherapist and mobilization commenced on the first post-operative day. Patient body temperature was routinely measured, three times daily (at 07:00, 13:00 and 19:00 h), by a trained nurse, using an infra-red tympanic thermometer and recorded on an observation chart which was filed in the patient's medical records. After TKR patients are routinely followed-up in the outpatient clinic for at least 5 years. Data collected from the medical records were: maximum body temperature recorded on each calendar day (pre-operative day, day of operation (day 0) and for 5 days (days 1–5) post-operatively or until discharge), whether an infection was diagnosed (any infection as an inpatient/infection of wound or prosthesis at any time during follow-up), whether a post-operative allogenic blood transfusion was given, preoperative and second post-operative day haemoglobin level, whether a urinary catheter was used, type of anaesthetic (general or regional) used and date of last outpatient follow-up. Data for pre-operative temperature (in 11 cases), day of operation temperature (in 5 cases), anaesthetic type (in one case) and second post-operative day haemoglobin level (in 63 cases) were missing from the medical records. For the purpose of our study, pyrexia is defined as a recorded body temperature of 38 °C or greater. Surgical wound infection, either superficial or deep, was diagnosed by the criteria suggested by the National Center for Infectious Diseases, Centers for Disease Control and Prevention, USA [4]. A urinary tract infection was defined as a positive urine culture that was treated with antibiotics. A respiratory tract infection was defined as either consolidation on chest X-ray and/or positive sputum culture that was treated with antibiotics. Statistical analysis of the association between pyrexia and infection, allogenic blood transfusion, haemoglobin loss, use of urinary catheter, rheumatoid arthritis, anaesthetic type, and previous pyrexia following TKR was by Fisher's Exact Test. Analysis of post-compared to pre-operative temperature was by Wilcoxon matched-pairs signed-ranks test. Statistical significance was set at p < 0.05. Data analysis and calculations were performed using the InStat statistical software package, version 3.0b for Macintosh (GraphPad Software, San Diego, CA, USA).
Fig. 1. Daily maximum temperature following TKR – means with 95% CI (box) and ranges (whiskers) shown.
Two patients developed a prosthetic infection. The first was a 60year-old woman with osteoarthritis who complained of increasing knee pain 6 months post TKR, in whom a Staphylococcus aureus prosthetic infection was diagnosed. She ultimately went on to have removal of the implant and arthrodesis of the knee. The second case was a 68-year-old man with rheumatoid arthritis who developed an inflamed knee 3 years after TKR. This was diagnosed as a deep S. aureus infection and was treated by arthroscopic debridement and long-term suppressive antibiotic therapy. There were no cases of DVT diagnosed in the study group. The maximum daily body temperatures are summarized in Table 2 and Figs. 1 and 2. It can be seen from Fig. 2 that there is a statistically significant increase in mean temperature from preoperation to post-operation, and that this increase remains significant through to 5 days post surgery (p < 0.0001). In 62/170 operations (36.5%) pyrexia was present at some point during the first 5 post-operative days. The daily prevalence of pyrexia is shown in Fig. 3, with the highest prevalence on day 1. The association between the various factors under investigation and the development of post-operative pyrexia are summarized in Table 3. There were no statistically significant associations,
3. Results Mean follow-up time was 19.2 months (range 2–72, interquartile range 6–28). There were 14 post-operative infections (infection rate 8.2%), all in different patients, detailed in Table 1. Table 2 Daily maximum temperature distribution (°C) of patients before and for 5 days after TKR Temperature
Pre-op
Day 0
Day 1
Day 2
Day 3
Day 4
Day 5
<36 36–36.9 37–37.9 38–38.9 ≥39 n % Pyrexial
0 146 13 0 0 159 0
11 109 35 10 0 165 6.1
7 53 79 31 0 170 18.2
3 47 92 27 1 170 16.5
1 71 81 17 0 170 10.0
1 92 64 12 0 169 7.1
0 95 61 3 0 159 1.9
Values indicate number of patients.
Fig. 2. Daily maximum temperature following TKR – means with 95% CI shown.
326
S. Ghosh et al. / The Knee 13 (2006) 324–327
Fig. 3. Daily prevalence of pyrexia ≥38 °C following TKR.
although haemoglobin loss of more than 3 g/dl was nearly significant (p = 0.075, odds ratio 2.167, 95% CI = 0.986–4.76). Of the 14 cases that developed a clinical infection, only four were pyrexial. There were 58 cases that developed pyrexia and had no clinical evidence of infection. Using the presence of pyrexia at any time as a diagnostic test for the development of infection gives a sensitivity of 0.286 (95% CI = 0.084–0.581), specificity of 0.628 (95% CI = 0.548–0.704) and positive predictive value of 0.065 (95% CI = 0.018–0.157).
4. Discussion Pyrexia in the first few days following many types of surgery, including joint replacement surgery, is a relatively common finding [1–3,5]. The physiological response to surgical trauma, disturbance of thermoregulation by anaesthesia, retained haematoma, atelectasis, pneumonia, urinary tract infection, superficial or deep wound infection, deep venous thrombosis (DVT) and response to blood transfusion have been suggested as some of the potential causes of a raised temperature in the TKR patient [1,2,6,7]. Amongst these many causes the overwhelming concern in the mind of the surgeon is that there may be a deep infection of the wound or prosthesis as this is associated with significant morbidity and mortality [8,9]. Due to this fear, TKR patients who develop a post-operative pyrexia are often subjected to a routine panel of investigations (chest X-ray, urine, blood and wound-swab cultures, serial full-blood-counts and inflammatory markers, Doppler ultrasound or venogram to exclude DVT) in an attempt to determine the cause of the fever. In our experience, applied blindly, these tests seldom produce useful diagnostic information. In addition, on development of a postoperative fever, additional antibiotics are often prescribed without firm evidence of infection being present, despite prolonged antibiotic use being potentially harmful [10]. The number of publications looking at pyrexia following TKR is quite limited, with some differing results. Kennedy et al. [1] looked at 90 TKR patients and found that 17%
developed a high pyrexia (≥ 39 °C), most commonly on the second post-operative day. They found significant independent associations between high pyrexia and both blood transfusion and fall in haematocrit. They found no association between pyrexia and the presence of an infective focus and concluded that pyrexia in the first few days following TKR is a normal physiological process. Guinn et al. [2] looked at 95 unilateral and 23 bilateral TKRs. They found a post-operative pyrexia (≥ 38 °C) rate of 66% for unilateral and 74% for bilateral TKRs. They found that patients who experienced a fever were significantly more likely to develop a complication in the immediate postoperative period, but there was no difference in the rate of long-term complications. No association between operative blood loss and pyrexia was found. They concluded that the rate of pyrexia after TKR is high and, based on the natural history of the fever, that atelectasis and DVT were the most likely causes. Shaw and Chung [7] examined 100 patients undergoing TKR plus 100 patients undergoing total hip replacement (THR). They found that the mean maximum temperature was highest on the first post-operative day and that 19% had a fever of ≥ 39 °C at some point. No association between blood loss or transfusion and pyrexia was found. Their conclusion was that post-operative fever after total joint arthroplasty is a normal physiological response and that a full work-up for fever alone is not justified. The policy in our unit is to use a four-dose antibiotic prophylaxis regimen and also to routinely use suction drains for all TKRs. Four doses of antibiotic, compared to three or less, has been shown to be effective in reducing infection rates following total hip replacement (THR) [11]. The routine use of wound drains is more controversial. A meta-analysis [12] has shown that, following TKR or THR, the use of drains was associated with a greater need for blood transfusion, yet made no difference to the rates of wound infection, haematoma, limb swelling, venous thrombosis or hospital-stay.
Table 3 Association between various factors and the development of a post-operative pyrexia following TKR
Infection (n = 170)
No Yes Transfusion (n = 170) No Yes Haemoglobin <3 g/dl loss (n = 107) ≥3 g/dl Urinary catheter (n = 170) No Yes Arthritis type (n = 170) Osteo Rheumatoid Anaesthetic type (n = 169) General Regional Previous pyrexia No after TKR (n = 49) Yes
No pyrexia Pyrexia
p
98 10 64 44 39 25 83 25 99 9 72 35 20 10
0.5786 NS 0.518 NS 0.075 NS 0.4421 NS 0.4268 NS 0.7378 NS 0.3774 NS
(58%) (6%) (38%) (26%) (36%) (23%) (49%) (15%) (58%) (5%) (43%) (21%) (41%) (20%)
58 (34%) 4 (2%) 40 (24%) 22 (13%) 18 (17%) 25 (23%) 51 (30%) 11 (6%) 54 (32%) 8 (5%) 40 (24%) 22 (13%) 10 (20%) 9 (18%)
Numbers of patients (and percentages) are shown. NS = not significant.
S. Ghosh et al. / The Knee 13 (2006) 324–327
The rate of deep infection of the prosthesis in our study group was 2/170 (1.2%) which, albeit with a shorter followup time, is comparable to other series [8,9]. Surprisingly, no case of DVT was diagnosed in our study group. This low rate compared to published series may be due to our universal use of chemical prophylaxis and compression stockings, early mobilization and the fact that only patients with significant clinical symptoms were investigated by venography [13]. Our results suggest that there is an element of postoperative hypothermia affecting the temperature of some patients on the day of surgery (temperature < 36 °C in 11/165 patients). Hypothermia in the first 12 h following TKR has been observed previously and the use of measures such as warming blankets and warmed intravenous infusions during surgery are recommended to minimize the problem [14]. In contrast to the studies of Kennedy et al and Shaw and Chung we found that the highest temperatures occurred on the second rather than first post-operative day and also that temperatures of > 39 °C were uncommon. We found no significant association between the development of a pyrexia and blood transfusion, use of urinary catheter, rheumatoid arthritis, anaesthetic type, and previous pyrexia following TKR. Blood transfusion is much less of a risk factor for the development of pyrexia since the change over to using leukocyte-depleted blood and this may partly explain why temperatures of > 39 °C were seldom seen in our cohort of patients [15]. Haemoglobin loss of more than 3 g/dl almost reached significance (p = 0.075) for an association with pyrexia. Haemoglobin loss, independent of subsequent transfusion, has been shown to be a factor associated with fever following TKR in the past and it has been suggested that the pyrexia is a response to haematoma [1,7]. The trend towards an association in our study, together with our universal use of suction drains, suggests that increased tissue trauma and its association with bleeding, rather than haematoma per se, may be the cause of the fever. Studies have suggested that a febrile reaction following TKR is at least partly a physiological response due to surgical site inflammation. Mehra et al. [16] have shown that, following TKR and in the absence of infection, local skin temperature and C-reactive protein level remain elevated for up to 18 weeks and 6 weeks respectively, indicating that a significant and persistent local inflammatory process is taking place. Andreas et al. [3] have shown that the levels of the endogenous pyrogen interleukin-6 in the serum and drain fluid of TKR patients was higher in those who were pyrexial after surgery, suggesting that this cytokine may be involved in the development of the fever. It is interesting to note that only four of the 14 patients in our study who developed an infection were pyrexial in the post-operative period. This may suggest that a lack of a pyrexia is a risk for infection, as the patient has not mounted an adequate immune response.
327
The results of our study support the hypothesis that most post-TKR fevers are physiological, as we found a high pyrexia rate (36%), a significant increase in mean body temperature throughout the 5-day post-operative period, and no correlation between pyrexia and the diagnosis of an infection. The low sensitivity, specificity and positivepredictive-value of pyrexia as a diagnostic test for the presence of infection shown in our study means that, on its own, the presence of pyrexia has little diagnostic value. In the absence of additional clinical findings, the presence of pyrexia in the first 5 days following TKR is usually a normal physiological finding and should not cause undue concern about the presence of infection.
References [1] Kennedy JG, Rodgers WB, Zurakowski D, Sullivan R, Griffin D, Beardsley W. et al. Pyrexia after total knee replacement, a cause for concern? Am J Orthop. 1997(Aug);26(8):549–52, 554. [2] Guinn S, Castro Jr FP, Garcia R, Barrack RL. Fever following total knee arthroplasty. Am J Knee Surg 1999 (Summer);12(3):161–4. [3] Andres BM, Taub DD, Gurkan I, Wenz JF. Postoperative fever after total knee arthroplasty: the role of cytokines. Clin Orthop Relat Res 2003 (Oct);415:221–31. [4] Guidline for prevention of surgical site infection. US Department of Health and Human services, 1999. www.cdc.gov/ncidod/dhqp/pdf/ guidelines/SSI.pdf (accessed March 2006). [5] Than P, Malovics I. Significance of postoperative fever after hip prosthesis implantation. Z Orthop Ihre Grenzgeb 2000 (Sep–Oct);138 (5):430–5. [6] Aburahma AF, Saiedy S. Deep vein thrombosis as probable cause of fever of unknown origin. W V Med J 1997 (Jan–Feb);93(1):368–70. [7] Shaw JA, Chung R. Febrile response after knee and hip arthroplasty. Clin Orthop Relat Res 1999 (Oct);367:181–9. [8] Lentino JR. Prosthetic joint infections: bane of orthopedists, challenge for infectious disease specialists. Clin Infect Dis 2003 (May 1);36 (9):1157–61. [9] Blom AW, Brown J, Taylor AH, Pattison G, Whitehouse S, Bannister GC. Infection after total knee arthroplasty. J Bone Joint Surg Br 2004 (Jul);86(5):688–91. [10] Hanssen AD, Osmon DR. The use of prophylactic antimicrobial agents during and after hip arthroplasty. Clin Orthop Relat Res 1999 (Dec); 369:124–38. [11] Engesaeter LB, Lie SA, Espehaug B, Furnes O, Vollset SE, Havelin LI. Antibiotic prophylaxis in total hip arthroplasty. Acta Orthop Scand 2003 (Dec);74(6):644–51. [12] Parker MJ, Roberts CP, Hay D. Closed suction drainage for hip and knee arthroplasty – a meta-analysis. J Bone Joint Surg Am 2004 (Jun);86-A(6):1146–52. [13] Warwick DJ, Whitehouse S. Symptomatic venous thromboembolism after total knee replacement. J Bone Joint Surg Br 1997 (Sep);79 (5):780–6. [14] Dalen T, Nilsson KG, Engstrom KG. Fever and autologous blood retransfusion after total knee arthroplasty: a prospective study of 40 autotransfusion events in 21 patients. Acta Orthop Scand 2002 (Jun); 73(3):321–5. [15] Zhao SM, Cheng XL, Hu J, Xiang GC, Zhang JS, Li RQ. Clinical assessment of preventing febrile nonhemolytic transfusion reaction by leukocyte-depleted blood transfusion. Zhongguo Shi Yan Xue Ye Xue Za Zhi 2002 (Dec);10(6):568–70. [16] Mehra A, Langkamer VG, Day A, Harris S, Spencer RF. C reactive protein and skin temperature post total knee replacement. Knee 2005 (August);12(4):297–300.