Since March 2014, there has been an outbreak of Ebola virus disease (EVD) in West Africa; the affected countries include Guinea, Liberia, Sierra Leone and, more recently, Nigeria. The cumulative number of confirmed EVD cases rose exponentially from May to June 2014.1 On 8 August 2014, the World Health Organization (WHO) declared the EVD outbreak a “Public Health Emergency of International Concern”, indicating that EVD is no longer a distant and confined issue, but a proximate and real threat.2

The Ebola virus was first discovered in 1976.3 4 It can be transmitted through direct contact with blood, secretions, and other body fluids or tissues of infected animals or persons.3 4 The incubation period of EVD ranges from 2 to 21 days,5 and a case fatality rate of 90% has been reported.5

Chemical pathology laboratory investigations are among the most basic and common tests requested for patients, particularly when intensive care is required, as would be expected for patients with EVD who are critically ill. Standard, contact, and droplet precautions have been recommended for the management of hospitalised patients with suspected EVD.6 While EVD is not normally transmitted by aerosol, there is a concern that the Ebola virus can remain infectious in laboratory-generated aerosol.7 8 9 Hence, stringent guidelines for laboratory personnel with respect to handling of laboratory specimens containing Ebola virus have been published.10 11 The WHO suggested in its interim guideline that “activities such as micro-pipetting and centrifugation can mechanically generate fine aerosols that might pose a risk of transmission of infection through inhalation as well as the risk of direct exposure”, and recommended that gown, gloves, eye-face protection, and particulate respirators such as the US National Institute for Occupational Safety and Health–certified N95 respirator should be used when laboratory personnel are performing activities such as aliquoting, centrifugation, and other procedures that may generate aerosol.10

Nowadays, most general chemistry tests are performed with analysers that aspirate specimens from primary blood collection tubes on which centrifugation has been performed. Flushing the instrumental parts with Triton X-100 (Dow Chemical Company, Midland [MI], US) or Clorox (Clorox, Oakland [CA], US) has been suggested for decontamination after analysis of specimens containing Ebola virus. However, such disinfection procedure may not achieve 100% inactivation of the virus and is likely to affect the chemical analysis. Therefore, general chemistry analysers are not suitable for analysing highly infectious specimens. Processing of these specimens without an adequate disinfection procedure will pose an occupational health hazard to laboratory workers. It has been suggested that specimens that potentially contain live Ebola virus should be processed in a class 2 biological safety cabinet following biosafety level 3 practices.6

With regard to inactivation of Ebola virus in blood specimens of patients, the heat inactivation procedure (incubation of serum or plasma specimens at 60°C for 60 minutes) has been reported to decrease viral titres in patients’ specimens.12 A report on the effect of the same heat inactivation procedure in the Hitachi 917 (Roche Diagnostics, Basel, Switzerland) and Bayer ACS 180 (Bayer Diagnostics, New York, US) biochemistry analysers has demonstrated minimal change in concentrations of sodium, potassium, urea, creatinine, glucose, urate, total bilirubin, amylase, and C-reactive protein, but significant reductions in concentrations of troponin, bicarbonate, total protein, albumin, total calcium, phosphate, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma-glutamyltransferase (GGT), and creatine kinase (CK).13 However, the report only stated the change in percentage, with no information provided for diagnostic utility (eg the correlation of the post-heat inactivation procedure result with the result obtained without the inactivation procedure, which would indicate retention of diagnostic information).13 14 Therefore, this study aimed to provide a statistical delineation of the heat inactivation procedure effects on more common general chemistry analytes using the AU5822 (Beckman Coulter Inc, Pasadena [CA], US) and troponin I using the Access 2 Immunoassay System (Beckman Coulter Inc).

Forty consecutive plasma specimens were obtained, anonymised, and divided into two aliquots. One aliquot of the specimens was analysed immediately, whereas the other aliquot was analysed after heat inactivation at 60°C for 60 minutes. In addition, as a pilot study, 20 specimens with elevated cardiac troponin I (>0.04 ng/mL, 99th percentile of the reference interval), together with 20 specimens with cardiac troponin I that were not elevated (below the 99th percentile) were analysed in the same manner. Common general chemistry tests—including sodium, potassium, chloride, urea, creatinine, glucose, total protein, albumin, total bilirubin, ALT, AST, ALP, GGT, amylase, total calcium, phosphate, magnesium, CK, and lactate dehydrogenase (LDH)—were performed on the AU5822 analyser and the troponin I test was done on the Access 2 Immunoassay System. The methodologies for the analytes are listed in Table 1 to allow laboratory staff using different analysers, but with similar methodologies, to adopt the results from the present study.

The heat inactivation procedure was performed according to a WHO guideline.12 Briefly, separated plasma aliquots were heated at 60°C for 60 minutes in a water bath while sealed in Eppendorf Tubes (Hamburg, Germany). The post-treatment specimens were then mixed inside the sealed tubes, allowed to settle, and analysed in the same manner as the untreated specimens.

Statistical tests were performed by MedCalc version 12.5 (MedCalc Software bvba, Ostend, Belgium). The pre- and post-treatment results were analysed with regard to normality by the Kolmogorov-Smirnov test, proportional change by Passing-Bablok regression, constant bias by paired t test, and maintenance of diagnostic value by Pearson correlation coefficient. When normality was not accepted by the Kolmogorov-Smirnov test, Wilcoxon test was used in place of paired t test, and Spearman rho in place of Pearson correlation coefficient. The P value of correlation was calculated for each analyte, with the linear regression accepted if the P value was <0.05, and linearity model accepted with the CUSUM (cumulative sum) test.

The effect of the heat inactivation procedure on an analyte was considered insignificant if the slope of the regression line was between 0.95 and 1.05, the linearity was preserved, and the correlation coefficient/rank correlation was ≥0.9. The constant bias was evaluated as to its statistical significance by paired t test or Wilcoxon test, and by consideration of clinical interpretation by two chemical pathologists. The effect of heat inactivation was considered interpretable if, despite a proportion bias, the linearity was preserved, and correlation coefficient/rank correlation was ≥0.9. If an analyte did not fulfil the above two criteria, the post-treatment measurement result was considered to have lost the diagnostic values.

A total of 20 chemical pathology tests were evaluated. Figure 1 shows the percentage change in concentrations of each analyte, and Table 2 shows the range of concentration, regression equation, and Pearson correlation coefficient of each analyte before and after heating. Among the analytes, nine tests (sodium, potassium, chloride, urea, creatinine, total calcium, phosphate, total protein, and glucose) were not significantly affected by the heat inactivation procedure and remained clinically interpretable.

Despite the proportional differences, analytical results for magnesium (15% mean increase), albumin (41% mean increase), bilirubin (8% mean decrease), amylase (27% mean decrease), and troponin I (76% mean decrease) were still interpretable, as the pre- and post-heat inactivation procedure results for these analytes were found to have significant correlation (P<0.0001 for the aforementioned analytes, with Pearson r or Spearman rho >0.9). These results can be interpreted with estimation from the significant proportional bias using the regression equation.

All the enzymes except for amylase (ALP, ALT, AST, CK, GGT, and LDH) were inactivated to a significant degree by the heat inactivation procedure, such that the results were considered uninterpretable. The Pearson r or Spearman rho values ranged from no significant correlation (P≥0.05) to 0.767, and most normality was rejected. The effect of the heat inactivation procedure on the analytes that were considered uninterpretable are shown in Figure 2.

The need for more stringent statistical considerations in the evaluation of disinfection procedures such as the heat inactivation procedure in the present study is evident. For example, the effect of the heat inactivation procedure on AST was quoted to have a mean reduction of 26% in one of the reports.13 An even lower average reduction of enzymatic activity (4.2%) was found in the present study although, despite the low reduction, diagnostic information was significantly destroyed (Spearman rho rank-order coefficient was only 0.767) after the heat inactivation procedure. In addition, consideration should be taken in interpretation of the clinical usefulness of post-heating results as some effects may be statistically significant but clinically insignificant in certain pathological conditions.

Our assays of total protein, total calcium, and phosphate showed no significant difference despite more adverse effects reported by Bhagat et al.13 This may be due to differences in analytical assays. Therefore, local laboratories should evaluate their assays.

It must be noted that the heat inactivation procedure is only one of the many facets of safe laboratory practice involving highly infectious materials. As centrifugation and micro-pipetting leads to aerosolisation,10 11 the use of gel separator tubes allows the heat inactivation procedure to be performed without micro-pipetting after centrifugation. As gel separator tubes have previously been reported to cause analytical interference in the analysis of biochemical analytes such as therapeutic drugs and steroid hormones,15 16 for hospitals that do not use gel separator tubes, verification of assay performance with gel separator tubes is necessary, although in the experience of the authors, the effect of gel separator tubes is usually minimal among general chemistry analytes.

Apart from general chemistry analytes, another panel of tests most commonly employed in the management of patients is blood gas analysis. It is, however, impossible to inactivate blood gas specimens using heat treatment as such treatment would invariably affect the acid-base, and the partial pressure of oxygen and carbon dioxide in blood, rendering the specimen unsuitable for analysis. Disinfection of routine blood gas analysers after the processing of infected specimens is also problematic; the glass electrode and membranes used for blood gas analysis are often not compatible with the high active chlorine content of bleach or the presence of surfactants in buffers used for disinfection purposes.

For blood gas analysis, the use of point-of-care analysers such as the i-STAT (Abbott Laboratories, Chicago [IL], US), whereby the test cartridge on which a blood specimen is applied is only connected to the analyser via electrical contacts, is seen as a viable alternative by the authors, as the test cartridge is single-use, can be disposed of safely, and the analyser can be cleaned and disinfected because of the vulnerable glass electrode, and the membranes on Clark- or Severinghaus-type electrodes (used for measurement of oxygen and carbon dioxide tensions) are not situated on the analyser properly.

Lastly, rather than performing the inactivation procedure in the laboratory, another option for analysis is the use of point-of-care analysers in an appropriate enclosure. Blood gas, electrolytes, and renal and liver function tests can all be performed in modern point-of-care analysers. In concordance with guidelines, it is recommended that the point-of-care analyser should be housed within a class 2 or above biosafety cabinet in a level 3 or above biosafety laboratory operating with appropriate precautions.11

We have presented the effect of the heat inactivation procedure on common biochemistry analytes, with statistical procedures applied to determine the diagnostic utility of the analyte concentrations. This serves to aid clinicians and laboratory staff in managing suspected and confirmed patients with EVD.

The authors would like to acknowledge Mr Kelvin Yu, Ms Kitty Soo, and Mr Philip Chiu for their kind assistance in performing the tests, and Ms Alision Lam for her kind assistance in the preparation of this manuscript.

1. World Health Organization. Disease Outbreak News (DONs) 2014 [12/8/2014]. Available from: http://www.who.int/csr/don/2014_08_11_ebola/en/. Accessed 12 Aug 2014.

2. IHR Emergency Committee Members and Advisers. WHO Statement on the Meeting of the International Health Regulations Emergency Committee Regarding the 2014 Ebola Outbreak in West Africa 2014 [12/8/2014]. Available from: http://www.who.int/mediacentre/news/statements/2014/ebola-20140808/en/. Accessed 12 Aug 2014.

3. Leroy EM, Kumulungui B, Pourrut X, et al. Fruit bats as reservoirs of Ebola virus. Nature 2005;438:575-6. Crossref

4. Wilson JA, Hevey M, Bakken R, et al. Epitopes involved in antibody-mediated protection from Ebola virus. Science 2000;287:1664-6. Crossref

5. Feldmann H, Geisbert TW. Ebola haemorrhagic fever. Lancet 2011;377:849-62. Crossref

6. Kortepeter MG, Martin JW, Rusnak JM, et al. Managing potential laboratory exposure to Ebola virus by using a patient biocontainment care unit. Emerg Infect Dis 2008;14:881-7. Crossref

7. Jaax N, Jahrling P, Geisbert T, et al. Transmission of Ebola virus (Zaire strain) to uninfected control monkeys in a biocontainment laboratory. Lancet 1995;346:1669-71. Crossref

8. Leffel EK, Reed DS. Marburg and Ebola viruses as aerosol threats. Biosecur Bioterror 2004;2:186-91. Crossref

9. Sewell DL. Laboratory-associated infections and biosafety. Clin Microbiol Rev 1995;8:389-405.

10. Interim infection prevention and control guidance for care of patients with suspected or confirmed Filovirus haemorrhagic fever in health-care settings, with focus on Ebola 2014. Available from: http://www.who.int/csr/resources/who-ipc-guidance-ebolafinal-09082014.pdf. Accessed 12 Aug 2014.

11. Interim guidance for specimen collection, transport, testing, and submission for patients with suspected infection with Ebola virus disease 2014 [12/8/2014]. Available from: http://www.cdc.gov/vhf/ebola/hcp/interim-guidance-specimencollection-submission-patients-suspected-infection-ebola.html. Accessed 12 Aug 2014.

12. WHO recommended guidelines for epidemic preparedness and response: Ebola haemorrhagic fever (EHF). Geneva: World Health Organization; 1997.

13. Bhagat CI, Lewer M, Prins A, Beilby J. Effects of heating plasma at 56 degrees C for 30 min and at 60 degrees C for 60 min on routine biochemistry analytes. Ann Clin Biochem 2000;37:802-4. Crossref

14. Hersberger M, Nusbaumer C, Scholer A, Knöpfli V, von Eckardstein A. Influence of practicable virus inactivation procedures on tests for frequently measured analytes in plasma. Clin Chem 2004;50:944-6. Crossref

15. Ferry JD, Collins S, Sykes E. Effect of serum volume and time of exposure to gel barrier tubes on results for progesterone by Roche Diagnostics Elecsys 2010. Clin Chem 1999;45:1574-5.

16. Bush V, Blennerhasset J, Wells A, Dasgupta A. Stability of therapeutic drugs in serum collected in vacutainer serum separator tubes containing a new gel (SST II). Ther Drug Monit 2001;23:259-62. Crossref