Usefulness of arterial blood gas analysis and D-dimer measurement in the assessment of pulmonary embolism after orthopedic surgery

J Orthop Sci (2006) 11:140–145 DOI 10.1007/s00776-005-0994-5

Original article Usefulness of arterial blood gas analysis and D-dimer measurement in the assessment of pulmonary embolism after orthopedic surgery Yasushi Oshima1,2, Shintaro Tachibana1, Yutaka Hirota1, Yusuke Takeda1, and Izuru Kitajima1 1 2

Department of Orthopaedic Surgery, Toranomon Hospital, Tokyo, Japan Department of Orthopaedic Surgery, Faculty of Medicine, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan

Abstract Background. Deep venous thrombosis (DVT) and pulmonary embolism (PE) after spinal or lower extremity surgery are well recognized as common complications. Since 1995 we have investigated the incidence of PE after orthopedic surgery using ventilation-perfusion (V/Q) lung scans, and the prevalence of PE was about 10%. With a view to detecting early-stage PE by simple examinations, we evaluated the use of both the blood gas analysis and the D-dimer measurement after spinal or lower extremity surgery. Methods. Altogether, 85 patients who underwent spinal or lower extremity surgery were eligible for the study. Pneumatic sequential leg compression devices (PSLCDs) were utilized continuously both intra- and postoperatively. Arterial blood gas analysis and D-dimer measurement were performed preand postoperatively on days 3 and 7. We set lung scan criteria as follows: postoperative decrease in Pao2 (∆Pao2) by ≥10 torr (group G), postoperative D-dimer of ≥10 µg/ml (group D), or both. Patients with the criteria went on to undergo lung scans, and PE was diagnosed by the existence of any mismatch between ventilation-perfusion (V/Q) lung scans. Results. A total of 44 (51.8%) patients met the lung scan criteria and underwent perfusion lung scans, 10 (11.7%) of whom were diagnosed as PE. In groups G and D, about 30% showed PE. Moreover, six (85.7%) of the seven patients with both criteria showed a significant increase (83.7%) in the prevalence of PE. Conclusions. Patients with the above criteria showed a high prevalence of PE. Moreover, 10 (11.7%) of the 85 patients were diagnosed as having PE, which corresponded to the prevalence in our former studies where lung scans were performed in all patients. The blood gas analysis and the Ddimer measurement may be utilized as the first screening examinations.

Offprint requests to: Y. Oshima Received: June 28, 2005 / Accepted: December 12, 2005

Introduction Deep venous thrombosis (DVT) and pulmonary embolism (PE) after lower extremity surgery are well recognized as common complications.1–3 Moser reported that 90% of PEs resulted from DVT, and 15% of DVTs could cause a PE.4 DVT and PE associated with spinal surgery have also been reported.5–7 Since 1995 we have investigated the incidence of PE after orthopedic surgery using ventilation-perfusion (V/ Q) lung scans.8–12 Formerly we performed lung scans in all patients with spinal or lower extremity surgery, and the prevalence of PE with pneumatic sequential leg compression devices (PSLCDs) both during and after surgery as prophylaxis was about 10%. However, lung scans are relatively expensive and technologydependent examinations, and more selective utilization of these tests has been advocated.13,14 The arterial blood gas analysis is a simple test and useful for excluding severe pulmonary diseases.14,15 Also, D-dimer levels have been reported to be related to onset of DVTs.16 Although neither test is diagnostic of PE by itself,13 each may be useful as the first screening examinations. The purpose of the current study was to evaluate the use of both blood gas analysis and the D-dimer measurement with a view to detecting early-stage PE after spinal surgery or lower extremity surgery. In this study, in terms of reducing the need for lung scans, we established lung scan criteria using postoperative Pao2 or Ddimer levels and performed lung scans selectively in patients who met the criteria. The effectiveness of the criteria was examined by comparing the PE prevalence with that in our former studies.

Y. Oshima et al.: Assessment of Pulmonary Embolism

Materials and methods Patients All patients who underwent spinal or lower extremity surgery at the authors’ institution between April 2003 and March 2004 were considered for the study. Among these patients, 20 were excluded. One patient refused consent for lung scans. Three patients had been receiving anticoagulation therapy. Two patients had undergone lower extremity surgery within 6 months. Three patients had a history of previous DVT. Three patients with trauma, three with rheumatoid arthritis, two with hemodialysis, and three with tumors were also eliminated because these diseases can affect D-dimer levels. Thus, 85 patients (80.9%) were eligible for the study. Altogether, 53 patients underwent spinal surgery, and 32 underwent lower extremity surgery. Informed consent was obtained from all the patients before entering the study. Complete demographic data were recorded for these patients, including age, gender, height, body weight, diagnosis, type of the procedure, operating time, and total blood loss. The patients consisted of 43 males and 42 females. The average age at operation was 62.2 years (range 16–88 years). The types of operations are summarized in Table 1. Diagnostic process Arterial blood gas analysis was performed pre- and postoperatively on days 3 and 7. D-dimer was measured pre- and postoperatively on days 3 and 7 using the latex agglutination reaction assessed by light scattering with an infrared beam (LPIA Ace D-dimer; Dia-Iatron, Tokyo, Japan). We set lung scan criteria as follows:

141

postoperative decrease in Pao2 (∆Pao2) by ≥10 torr (group G) or postoperative D-dimer of ≥10 µg/ml (group D). For cutoff values of Pao2 and D-dimers, we referred to past reports in which the relation between these parameters and the prevalence of PE or DVT was investigated.13–16 Patients who met the criteria went on to perfusion lung scans. If there were any defects in the perfusion lung scans, ventilation lung scans were additionally performed. PE was diagnosed by the existence of any mismatch between ventilation-perfusion (V/Q) lung scans (Fig. 1). All scans were evaluated by radiologists in a blinded manner.

Table 1. Operative procedures Level/joint Spine Cervical

Thoracic Lumbar

Lower extremities Knee Hip

Procedure

No. of patients

ADF Laminoplasty ADF + laminoplasty Laminectomy PLF Laminectomy Fenestration Disectomy PLF PLIF Laminectomy

3 21 1 1 1 2 5 9 1 7 2

TKA THA RAO

11 17 4

ADF, anterior decompression and fusion; PLF, posterolateral fusion; PLIF, posterior lumbar interbody fusion; TKA, total knee arthroplasty; THA, total hip arthroplasty; RAO, rotational acetabular osteotomy

Fig. 1. Flow chart summarizing the diagnostic process. The number of patients involved and the ratios are also shown

142

Y. Oshima et al.: Assessment of Pulmonary Embolism

Table 2. Statistical analysis of risk factors for PE between postoperative PE-positive and PE-negative groups Parameter Age (years) Postoperative bed rest (days) Body mass index Operating time (minutes) Total blood loss (g)

PE positive (n = 10) 73.9 ± 4.5 ± 25.7 ± 201.3 ± 263.2 ±

7.8 2.7 3.4 118.8 295.6

PE negative (n = 34) 64.1 4.9 24.1 200.2 341.9

± 15.4 ± 4.6 ± 3.5 ± 94.0 ± 336.6

P* 0.03** 0.90 0.14 0.48 0.43

Results are means ± SD PE, pulmonary embolism * Mann-Whitney test ** Statistically significant

Prophylaxis Intraoperatively, all but the patients with total knee arthroplasty (TKA) were given pneumatic sequential leg compression devices (PSLCDs) bilaterally. Only TKA patients had PSLCDs applied to the unaffected side. Postoperatively, the patients had been required to put on both the devices and elastic stockings until they were transferred to wheelchairs or started ambulating. No anticoagulation therapy was performed. Statistical analysis The Mann-Whitney test, paired t-test, and Wilcoxon signed-ranks test were used for the numerical data. Statistical analysis of the occurrence of specific events in the two study groups was performed using Fisher’s 2 × 2 exact test. The differences between the two groups were considered statistically significant when P < 0.05. All reported P values were two-sided.

Table 3. Comparison between PE-positive and PE-negative groups Parameter Group G (n = 28) No. of patients PaO2 (torr) Preoperative Day 3 Day 7 PaCO2 (torr) Preoperative Day 3 Day 7 Group D (n = 23) No. of patients D-dimer (µg/ml) Preoperative Day 3 Day 7

PE-positive

PE-negative

P*

9

19

83.1 ± 8.4 72.5 ± 15.8 83.9 ± 14.1

88.6 ± 9.0 75.6 ± 9.5 82.3 ± 13.9

0.08 0.24 0.60

41.0 ± 3.4 39.0 ± 5.0 38.6 ± 2.3

40.1 ± 3.4 41.6 ± 3.5 39.0 ± 5.0

0.23 0.09 0.62

7

16

2.2 ± 2.0 6.5 ± 3.0 15.8 ± 4.4

1.8 ± 1.6 7.7 ± 4.9 12.5 ± 2.6

0.59 0.69 0.06

Results are means ± SD. * Mann-Whitney test

PE prevalence in groups G and D Results Overall prevalence of PE Among the 85 patients, 44 (51.8%) met the lung scan criteria and underwent perfusion lung scans; 20 (23.5%) of them went on to ventilation lung scans owing to defects in the perfusion lung scans, and 10 (11.7%) were found to have a mismatch between the V/Q lung scans, which was diagnostic of PE (Fig. 1). Two patients (2.1%) showed a clinical PE, neither of which was fatal. As risk factors for PE, statistical analysis between PEpositive and PE-negative patients indicated that only age had significant correlation with PE prevalence (Table 2). The interval between surgery and perfusion or ventilation lung scans ranged from 3 to 10 days (average 7.7 days) or 4 to 15 days (average 10.1 days), respectively.

A total of 28 patients were included in group G and 23 in group D. Seven were included in both groups. Of the 28 patients in group G, PE was detected in nine (32.1%). Postoperative Pao2 and Paco2 showed no significance between PE-positive and PE-negative patients in this group (Table 3). For group D, 7 (30.4%) of the 23 patients had a PE, whereas postoperative D-dimer measurement showed no significance in this group (Table 3). Altogether, 21 patients belonged only to group G but not to group D; 3 (14.3%) of them had a PE. Similarly, of the 16 patients only in group D (not in group G), 1 (6.3%) had a PE. On the other hand, six (85.7%) of the seven patients who belonged to both groups G and D had a PE, which was statistically significant (Fig. 2). Comparison of days 3 and 7 Overall, postoperative Pao2 levels were significantly lower on day 3 than those on day 7, whereas postopera-

Y. Oshima et al.: Assessment of Pulmonary Embolism

143

Table 4. Overall average of Pao2, Paco2, and D-dimer levels Parameter Pao2 (torr) Paco2 (torr) D-dimer (µg/ml)

Preoperative

Day 3

Day 7

P*

86.3 ± 9.9 40.5 ± 3.7 1.3 ± 2.1

81.4 ± 12.6 39.7 ± 3.7 3.6 ± 3.5

87.5 ± 13.9 38.8 ± 4.2 6.4 ± 5.2

0.02**† 0.57** <0.00001**†

* Comparison between days 3 and 7 ** Wilcoxon signed-ranks test *** Paired t-test † Significant

Table 5. Lung scan criteria and surgical sites Location Spine (n = 53) Lower extremity (n = 32) a

Fig. 2. Combination of groups G and D resulted in a significantly increased incidence of pulmonary embolism (*Fisher’s 2 × 2 exact test)

tive D-dimer levels were significantly higher on day 7 (Table 4). Of the nine PE-positive patients in group G, eight met the criterion on day 3, and the rest (one patient) on day 7. On the other hand, all PE-positive patients in group D met the criterion only on day 7. PE prevalence and surgical site In the spinal surgery group, 21 of the 53 patients met the criteria, and 4 (19.0%) of the 21 patients with the criteria had a PE. In the lower extremity surgery group, 23 of the 32 patients met the criteria, and a PE was found in 6 (26.1%) (Table 5). There were no significant differences in the PE prevalence between these two groups (Fisher’s 2 × 2 exact test).

Discussion Lower extremity or spinal surgery is known to be associated with a high prevalence of PE, which has been well documented in Western countries and also in Japan recently.1,2,7 Furthermore, PE is also generating growing interest as an “economy-class syndrome.” Investigators have explored numerous diagnostic tests and procedures by which clinicians can identify PE without resorting to the gold standard of pulmonary

Criteria (+)

PEa

21 23

4 (19.0%) 6 (26.1%)

PE prevalence among patients who met the lung scan criteria

angiography, which is rather invasive.17 Clinical probability assessment by lung scans has been correlated with the presence of PE, as shown in the Prospective Investigation on Pulmonary Embolism Diagnosis (PIOPED) study.18 Harris et al. investigated the prevalence of PE after total hip arthroplasty (THA) using lung scans, and reported that the prevalence was 23%, 17% of which were clinical.2 McCardel et al. also reported that the prevalence with anticoagulation therapy was 12.6%, and clinical PE was present in 1.9% according to lung scans.3 Similarly, thromboembolic disease is a significant complication in patients who undergo spinal surgery. Oda et al. reported that DVT occurred in 15.5% of patients with spinal surgery.7 As for PE prevalence after spinal surgery, many are reported. Dearborn et al. investigated 318 patients with spinal surgery and reported the prevalence of clinical PE to be 2.2% without any prophylaxis.6 Wood et al. published a study of 136 patients in whom mechanical prophylaxis was used, and reported that the prevalence of clinical PE was 0.7%.19 Since 1995 we have investigated the incidence of PE after orthopedic surgery using lung scans (Table 6). Formerly, we performed perfusion and lung scans in all patients with spinal or lower extremity surgery. The incidence of PE after THA without prophylaxis was 25%,8 which decreased slightly when using PSLCDs postoperatively.9 Postoperative recumbency for more than 7 days was found to be a risk factor. Thereafter, PSLCDs were used both intra- and postoperatively, and the prevalence decreased to about 10%.10,11 The prevalence of PE after spinal surgery was 10% in our past study.12 For any kind of surgery, certain patient characteristics have been identified as risk factors for venous thromboembolism, such as increasing age, prolonged

144

Y. Oshima et al.: Assessment of Pulmonary Embolism

Table 6. Incidence of PE in our former studies No. of patients Year 1999 2000 2003 2003 2004 Current study a

Surgical site or procedure THA Lower extremity Lower extremity Spine THA and TKA Spine and lower extremity

Surgery

Lung scan

PE

Incidence (%)

No. of patients with clinical PE

Prophylaxisa

32 69 63 50 68 85

32 69 63 50 68 44

8 14 6 5 7 10

25.0 20.3 9.5 10.0 10.2 11.7

1 0 0 1 0 2

None Postoperative Intra- and postoperative Intra- and postoperative Intra- and postoperative Intra- and postoperative

Pneumatic sequential leg compression devices

bed rest, obesity, and previous venous thromboembolism.20 Among these factors, only age showed significance in the current study. It would be preferable to predict or even detect a PE before its onset by simple examinations, but none of the numerous diagnostic examinations is valuable as a single diagnostic tool.13 Blood gas analysis is a simple test and its disturbances are commonly observed with both PEs and a number of other respiratory disorders. Alone, however, it is not diagnostic. Moreover, most PEs after orthopedic surgery found in our former studies occurred subclinically. However, it is reported that normal blood gas values can exclude most clinical PEs and may be useful in combination with clinical symptoms or other examinations.15 D-dimer is a product of fibrinolysis that is released in the blood as a result of dissolved clots and thrombus, and it has been reported to be related to the onset of DVT. Shiota et al. retrospectively investigated D-dimer levels using a latex photometric immunoassay system (LPIA) in patients after THA and TKA and reported that a cutoff value of 10 µg/ml on day 7 had a high degree of association with DVT with more than 90% specificity and sensitivity.16 D-dimer may be an excellent parameter for diagnosing DVT, but there have been few reports of a relation between postoperative D-dimer levels and the incidence of PE. The prevalence of PE was about 30% in groups G and D, which means about 70% were false-positive. However, the ratio was about three times as high as that of our past reports where V/Q lung scans were performed in all patients. Accordingly, as a single parameter, each of them may help detect PEs to some extent. Moreover, if both criteria were met (i.e., if ∆Pao2 was more than 10 torr and the postoperative D-dimer level was more than 10 µg/ml), the PE prevalence was 85.7%. On the other hand, if only one of them was abnormal, the PE prevalence was significantly low (Fig. 2). Consequently, patients who meet both criteria are at extremely high risk for PE.

In this study, of all 85 patients, 44 (51.8%) underwent pulmonary lung scanning, and 10 (11.7%) were diagnosed as having a PE. Unfortunately, we did not carry out lung scans in patients who did not meet the criteria. However, the PE prevalence in such patients was thought to be low because the PE prevalence in this study was almost equivalent to that in our former studies where pulmonary scans were performed in all patients (Table 6). This indicates that patients who do not meet the criteria are at low risk for PE, and that patients only with the criteria should be subjected to pulmonary scanning. That is, blood gas analysis and D-dimer measurement can be utilized for detecting early-stage PE as the first screening examinations before performing lung scans. Although we performed blood gas analysis and Ddimer measurement postoperatively on days 3 and 7, postoperative Pao2 levels were significantly lower on day 3 than those on day 7, whereas postoperative Ddimer levels were significantly higher on day 7 (Table 4). Indeed, all PE-positive patients in group D met the criterion only on day 7. For group G, eight out of the nine PE-positive patients met the criterion on day 3. These findings indicate that the blood gas analysis on day 3 and D-dimer measurement on day 7 might have been enough for the first screening examinations. As for PE and surgical sites, more patients in the lower extremity group met the criteria (not significant, Fisher’s 2 × 2 exact test), but the PE prevalence among patients with the criteria was similar for spinal surgery group and the lower extremity group (Fisher’s 2 × 2 exact test) (Table 5). Accordingly, the criteria are not supposed to depend on the surgical site or the procedure.

Conclusions Selective V/Q lung scans employed in this study with the above criteria showed that the 30% PE prevalence

Y. Oshima et al.: Assessment of Pulmonary Embolism

of patients who met the criteria nearly tripled that of our past reports. Moreover, patients who met both criteria showed a significant increase in the prevalence of PE. Also, 11.7% of PE detected in this study corresponded to the prevalence in our former studies, where lung scans were performed in all patients. Although the design of this study has a limitation in that we did not carry out lung scans in all patients, we propose that blood gas analysis and D-dimer measurement be inserted into the diagnostic protocol for PE as the first screening examinations.

References 1. Fujita S, Hirota S, Oda T, Kato Y, Tsukamoto Y, Fuji T. Deep venous thrombosis after total hip or total knee arthroplasty in patients in Japan. Clin Orthop 2000;375:168–74. 2. Harris WH, McKusick K, Athanasoulis CA, Waltman AC, Strauss HW. Detection of pulmonary emboli after total hip replacement using serial C15O2 pulmonary scans. J Bone Joint Surg Am 1984;66:1388–93. 3. McCardel BR, Lachiewicz PF, Jones K. Aspirin prophylaxis and surveillance of pulmonary embolism and deep vein thrombosis in total hip arthroplasty. J Arthroplasty 1990;5:181–5. 4. Moser, KM. Pulmonary embolism. Am Rev Respir Dis 1977;115:829–52. 5. Brambilla S, Ruosi C, La Maida GA, Caserta S. Prevention of venous thromboembolism in spinal surgery. Eur Spine J 2004;13:1–8. 6. Dearborn JT, Hu SS, Tribus CB, Bradford DS. Thromboembolic complications after major thoracolumbar spine surgery. Spine 1999;24:1471–6. 7. Oda T, Fuji T, Kato Y, Fujita S, Kanemitsu N. Deep venous thrombosis after posterior spinal surgery. Spine 2000;25:2962–7. 8. Kitajima I, Tachibana S, Hirota Y, Torio T. The incidence of pulmonary embolism following total hip arthroplasty. Seikeigeka (Orthopedic Surgery) 1999;50:1287–90 (in Japanese). 9. Hirota Y, Tachibana S. The incidence of pulmonary embolism following lower extremity surgery. Kansetsugeka (Joint Surgery) 2000;19:1377–82 (in Japanese).

145 10. Kawai T, Tachibana S, Hirota Y, Takeda Y, Nishimagi T, Ito T. Prophylaxis against pulmonary embolism using pneumatic compression therapy. Higashinihonseisaikaishi (Journal of the Eastern Japan Association of Orthopaedics and Traumatology) 2003;15:647–50 (in Japanese). 11. Hirota Y, Tachibana S, Kitajima I. The incidence and risk factors of pulmonary embolism following total hip and knee arthroplasty. Kotsukansetsujintai (Journal of Musculoskeletal system) 2004;17:607–11 (in Japanese). 12. Ito T, Takeda Y, Kawai T, Nakamichi K, Nishimagi T, Shiraki Y, et al. The incidence of pulmonary embolism after spinal surgery. Nippon Seikeigekagakkai Zasshi (J Jpn Orthop Assoc) 2003;77:190 (in Japanese). 13. Egermayer P, Town GI, Turner JG, Heaton DC, Mee AL, Beard MEJ. Usefulness of D-dimer, blood gas, and respiratory rate measurements for excluding pulmonary embolism. Thorax 1998;53:830–4. 14. McFarlane MJ, Imperiale TF. Use of the alveolar-arterial oxygen gradient in the diagnosis of pulmonary embolism. Am J Med 1994;96:57–62. 15. Stein PD, Goldhaber SZ, Henry JW, Miller AC. Arterial blood gas analysis in the assessment of suspected acute pulmonary embolism. Chest 1996;109:78–81. 16. Shiota N, Sato T, Nishida K, Matsuo M, Takahara Y, Mitani S, et al. Changes in LPIA D-dimer levels after total hip or knee arthroplasty relevant to deep-vein thrombosis diagnosed by bilateral ascending venography. J Orthop Sci 2002;7:444–50. 17. Stein PD, Athanasoulis C, Alavi A, Greenspan RH, Hales CA, Saltzman HA, et al. Complications and validity of pulmonary angiography in acute pulmonary embolism. Circulation 1992;85:462–8. 18. Prospective Investigation of Pulmonary Embolism Diagnosis Investigators. Value of the ventilation/perfusion scan in acute pulmonary embolism: results of the prospective investigation of pulmonary embolism diagnosis. JAMA 1990;263:2753–9. 19. Wood KB, Kos PB, Abnet JK, Ista C. Prevention of deep-vein thrombosis after major spinal surgery: a comparison study of external devices. J Spinal Disord 1997;10:209–14. 20. Tubiana R, Duparc J. Prevention of thrombo-embolic complications in orthopaedic and accident surgery. J Bone Joint Surg Br 1961;43:7–15.