Multiplex PCR for the Detection of Mycoplasma fermentans, M. hominism, and M. penetrans in Patients with Chronic Fatigue Syndrome, Fibromyalgia, Rheumatoid Arthritis, and Gulf War Syndrome

Journal of Chronic Fatigue Syndrome (The Haworth Medical Press, an imprint of the Haworth Press, Inc.) Vol. 5, No. 3/4, 1999, pp. 187-197

Aristo Vojdani, PhD, MT and Al Robert Franco
Aristo Vojdani is affiliated with the Immunosciences Laboratory Inc., Beverly Hills, CA, and also with the Department of Medicine, Drew University School of Medicine and Science, Los Angeles, CA.

Al Robert Franco is affiliated with The Arthritis Center of Riverside, CA.

Address correspondence to: Aristo Vojdani, PhD, MT, 8730 Wilshire Boulevard, Suite 305, Beverly Hills, CA 90211 (E-Mail: immunsci@ix.netcom.com).

[Haworth c-indexing entry note]: “Multiplex PCR for the Detection of Mycoplasma fermentans, M. hominis, and M. penetrans in Patients with Chronic Fatigue Syndrome, Fibromyalgia, Rheumatoid Arthritis and Gulf War Syndrome.” Vojdani, Aristo, and Al Robert Franco. Co-published simultaneously in Journal of Chronic Fatigue Syndrome (The Haworth Medical Press, an imprint of the Haworth Press, Inc.) Vol. 5, No. 3/4, 1999, pp. 187-197; and: Chronic Fatigue Syndrome: Advances in Epidemiologic, Clinical, and Basic Science Research (ed: Roberto Patarca-Montero) The Haworth Medical Press, an imprint of The Haworth Press, Inc., 1999, pp. 187-197. Single or multiple copies of this article are available for a fee from The Haworth Document Delivery Service [1-800-342-9678, 9:00 a.m. – 5:00 p.m. (EST). E-mail address: getinfo@haworthpressinc.com].

© 1999 by The Haworth Press, Inc. All rights reserved.

CHRONIC FATIGUE SYNDROME

SUMMARY. A multiplex polymerase chain reaction (PCR) was used to detect mycoplasma infection in human DNA samples of patients with CFS and related illnesses. One set of oligonucleotide primers which are specific for a highly conserved region among all members of genus Mycoplasma along with three other primer sets which are specific for Mycoplasma fermentans, M. Hominis, and M. penetrans species were used in this assay. The sensitivity of detection was determined by adding known mycoplasma DNA copy numbers to 1 µg of genomic DNA from healthy subjects. Each sample was subjected to 40 cycles of amplification. The detection level was determined to be 7, 7, 9, and 15 mycoplasma DNA copies per µg of human genomic DNA for M. genus, M. fermentans, M. hominis, and M. penetrans, respectively. The assay was applied to DNA extracted from the PBMCs of individuals suffering from chronic fatigue syndrome (CFS) (n=100), fibromyalgia (FMS) (n=40), rheumatoid arthritis (RA) (n=60), and gulf war syndrome (GWS) (n=60) and compared to age- and sex-matched healthy individuals (n=160). The percentage of M. genus infection detected in CFS, FMS, RA, and GWS was 52, 54, 49, and 55%, respectively. M. fermentans was detected in 32, 35, 23, and 36%, M. hominis was detected in 9,8,11, and 5%, and M. penetrans was detected in 6, 4, 7, and 3% of CFS, FMS, RA, and GWS patients respectively. M. genus, M. fermentans, M. hominis, and M. penetrans were detected in 15, 8, 3, and 2% of healthy matched controls. This assay provides a rapid and cost efficient procedure to screen clinical samples for the presence of three potentially pathogenic species of Mycoplasma with a high level of sensitivity and specificity. [Article copies available for a fee from The Haworth Document Delivery Service: 1-800-342-9678. E-mail address: getinfo@haworthpressinc.com ]

KEYWORDS. Multiplex polymerase chain reaction (PCR), Mycoplasma fermentans, M. hominis, M. penetrans, and chronic fatigue syndrome, fibromyalgia, rheumatoid arthritis, and gulf war syndrome.

INTRODUCTION

Mycoplasmas are prokaryotes belonging to the class Mollicutes. They are the smallest free-living self-replicating bacteria known (1). Mycoplasmas have no cell wall and a very limited genome of between 600 to 1500 kbp. The fact that they have such a limited genome makes them highly dependent upon their host for survival, much like viruses (1,2,11). Recent studies have implicated certain mycoplasma species as possible cofactors in a number of clinical conditions. Many different investigators have isolated Mycoplasma fermentans, M. hominis, and M. penetrans from individuals suffering from a variety of diseases including chronic fatigue syndrome, rheumatoid arthritis, and gulf war syndrome (3-5). Although the role that mycoplasmas may play in these clinical conditions is a subject of many debates, it does raise intrigue as to their possible involvement in chronic illnesses such as these and others with related or overlapping symptoms.

It has been well documented that people who suffer from fibromyalgia exhibit many of the same symptoms found in CFS (6,7). These two illnesses are so similar that for years many medical practitioners have considered them as the same condition. They are still regarded as closely associated in the scientific literature with only the exception of a few distinction criteria. Other research has stated the uncanny similarities between chronic fatigue syndrome and gulf war syndrome (8). Patients suffering from rheumatoid arthritis also exhibit certain symptoms characteristic to each illness (9). Although RA exhibits a narrower spectrum of clinical symptoms than CFS, FMS, and GWS, it does exhibit a significant overlap of symptoms found in each condition. The fact that these four illnesses have such a high degree of similarities led us to investigate the possibility that mycoplasma infection may be a common link to each condition. A successful clinical practice greatly depends on rapid and reliable detection techniques. Mycoplasma detection by culture is laborious and may take as long as five weeks to generate results that even then may be inconclusive or inaccurate due to the fastidious nature of mycoplasmas (10-12). Enzyme-linked immunosorbent assays (ELISA) to detect antibodies directed against mycoplasma species may also be used. Although it has a rapid turn-around time, the assay may lack sensitivity and specificity (13). More recently, molecular techniques using DNA probes and the polymerase chain reaction (PCR) have been developed in an attempt to improve upon current mycoplasma detection assays (14-16). Many researchers have used conventional PCR as a basis to develop multiple species detection assays called multiplex PCR (17,18). Multiplex PCR allows for the simultaneous detection and differentiation of multiple species in a single reaction with a level of sensitivity and specificity unmatched by any other diagnostic procedure (18). This assay enables clinicians to rapidly screen individuals for multiple species of potentially pathogenic mycoplasma while eliminating the cost and labor of individual assays to achieve the same results. The present study describes the use of a multiplex PCR for the detection of mycoplasmas at the genus level and the simultaneous detection and differentiation of Mycoplasma fermentans, M. hominis, and M. penetrans on clinical samples of patients with related illnesses.

MATERIALS AND METHODS

Clinical Specimens

A total of 100, 60, 40, and 40 CFS, RA, FMS and GWS patient samples respectively, were chosen for this study from various clinics throughout the country. The ages and the sex of the patients were matched with 160 healthy control subjects. Each patient had been ill for 1-5 years, and any other conditions that may cause CFS-like symptoms excluded individuals from the study. All blood samples were obtained under identical aseptic conditions to eliminate possible contamination and variation between samples.

DNA Isolation

Whole blood was collected in yellow top tubes containing acid citrate dextrose (ACD) solution A (Becton Dickinson, Franklin Lakes, NJ). The whole blood was gently layered over Histopaque (Sigma, St. Louis, MO) and centrifuged at 2000 rpm for 30 minutes. The peripheral blood mononuclear cells (PBMC were collected and washed twice in PBS (pH 7.40). DNA from the PBMC and cell lines were extracted by the same method. The cells were treated with 10 mM TRIS-HCI (pH 8.0), 1 mM EDTA (pH 8.0) (TE), 1% SDS with 20 µg ml-1 proteinase K at 55oC for 2 hours. Phenol-chloroform-isoamyl alcohol (25:24:1) was added to each sample to extract the DNA. The DNA was precipitated with 1/10 volume of 3 M sodium acetate and 2 volumes of absolute ethanol and incubated at -20oC overnight. The samples were subjected to centrifugation at 14,000 g for 20 minutes and then washed in 70% ethanol. A centrivap concentrator (Labconco, Kansas City, MO) was used for 12 minutes at 60oC to dry the pellets. The DNA pellets were resuspended in 100 µl of sterile TE. The concentration and purity were determined spectrophotometrically by measuring the absorbance at 260 and 280 nm. Human genomic DNA sample concentrations were all standardized at 0.2 mg ml-1 and then stored at -20oC until used.

Primer Selection

The four sets of oligonucleotide primers listed in Table 1 were selected based on their ability to efficiently amplify specific target sequences under the same reaction condition.

Table 1. Oligonucleotide primers used in the multiplex PCR
Mycoplasma
Primer

Amplification Product
genusa 5′ GGG AGC AAA CAG GAT TAG ATA CCC T 3′
5′ TGC ACC ATC TGT CAC TCT GTT AAC CTC 3′
280 bp
M. hominisb 5′ ATA CAT CGA TGT CGA GCG AG 3′
5′ CAT CTT TTA GTG GCG CCT TAC 3′
170 bp
M. termentansc 5′ GGA CTA TTG TCT AAA CAA TTT CCC 3′
5′ GGT TAT TCG ATT TCT AAA TCG CCT 3′
206 bp
M. penetransd 5′ CAT GCA AGT CGG ACG AAG CA 3′
5′ AGC ATT TCC TCT TCT TAC AA 3′
407 bp
a(19), b(1), c(2), d(1)

Multiplex PCR

The multiplex PCR reaction components and cycling parameters were determined through a number of experimental amplifications. We experimented with different DNA and reagent concentrations, annealing temperatures, and mycoplasma DNA copy numbers in various combinations. The optimized reaction was carried out in a final volume of 100 µl and each reaction mixture contained 10 mM TRIS-HCI (pH 8.3), 50 mM KCI, 1.5 mM MgCl2, 200 µM of each dATP, dCTP, dGTP, and dTTP, 50 DNA. The DNA amplification was performed in a GENE AMP 9600 thermal cycler (Perkin-Elmer, Norwalk, CT). The reaction parameters consisted of an initial 3 minute denaturation step at 95oC followed by 40 amplification cycles which consisted of a denaturation step at 94oC for 45 seconds, an annealing step at 55oC for 1 minute, and an extension step at 72oC for 2 minutes. The final cycle was followed by an additional extension step at 72oC for 10 minutes. A volume of 20 µl from each reaction was separated on a 1.8% agarose gel (FMC, Rockland, MD) stained with 0.5 µg ml-1 of ethidium bromide (Sigma, St. Louis, MO). The horizontal gels were submerged in TBE buffer (90 mM TRIS-borate, and 2 mM EDTA) and run at 80 V. The gels were exposed to a U.V. transilluminator to visualize the amplified products.

Multiplex PCR Sensitivity and Specificity

To determine the sensitivity level of this assay, PBMCs were isolated from a healthy individual who was negative for all mycoplasma species PCR. The cells were stained with 0.4% Trypan Blue (Sigma, St. Louis, MO) and counted using a haemocytometer. Serial dilutions of known quantities of purified mycoplasma DNA were added to 1 µg of human genomic DNA and subjected to the amplification procedure. We converted the added DNA quantities into bacterial cell copy numbers using the genome size of each mycoplasma species. This procedure enabled us to maintain control over the number of DNA templates that were added to the reaction, therefore verifying the detection limit of the assay.

We also introduced various concentrations and combinations of mycoplasma DNA copies from each species into the same reaction ranging from 1 to 50 copies per µg of human genomic DNA. This was done to determine the ability of the multiplex PCR to detect multiple species in the same reaction and the level of sensitivity at which it does so. The lowest number of mycoplasma cell copies where all species were clearly visible using agarose gel electrophoresis was considered to be the detection limit for the assay.

The reaction specificity was checked for the possibilities of cross-reactions with other mycoplasma species and closely related Gram-positive bacteria. The reaction fidelity was assessed by adding 100 mg of purified DNA from M. genitalium (49123), M. orale (23714), M. pirum (25960), M. pneumoniae (15531), M. arthritidis (35943), C. innocuum (14501), C. ramosum (25582), B. subtilis (6051), and E. coli (11775) (ATCC, Rockville, MD) to the reaction mixture in the absence and presence of negative control human genomic DNA and subjecting the samples to amplification under the conditions previously described.

RESULTS

Each mycoplasma species that was targeted by this assay was detectable at different bacterial cell copy numbers present among 1 µg of human genomic DNA. The Mycoplasma genus and M. fermentans primer sets had a detection limit of 7 mycoplasma DNA copies per µg of human DNA, whereas the M. hominis primer set was slightly less sensitive with the ability to detect 9 mycoplasma DNA copies per µg of human DNA. The M. penetrans primer set had the lowest sensitivity overall with the ability to detect 15 DNA copies of that species in the presence of 1 µg of human genomic DNA. The mycoplasma genomic DNA number detection limits were established and confirmed with the use of purified mycoplasma DNA. There was no cross-reaction with any of the species-specific primer sets when presented with control DNA from other mycoplasma species. The Mycoplasma genus primer set was able to amplify the predicted 280-bp region of each different mycoplasma species that was added to the reaction, and did not react with any of the non-mycoplasma DNA controls. Also, when the multiplex PCR was applied to clinical samples, the assay was able to detect each target sequence without any cross-reaction or interference from background DNA. The amplified products from actual clinical samples that consisted of M. penetrans (407 bp, M. genus (280 bp), M. fermentans (206 bp), and M. hominis (170 bp) were clearly detectable when visualized by agarose gel electrophoresis (Figure 1).

Figure 1. Agarose gel electrophoresis of multiplex polymerase chain reaction (PCR) amplified products generated from CFS patient DNA samples. Lane 1 and 11 are DNA size markers. Lane 2 shows 280 bp M. genus amplification product. Lane 3 shows M. genus and 206 bp M. fermentans. Lane 4 shows M. genus and 170 bp M. hominis. Lane 5 shows M. genus and 470 bp M. penetrans. Lane 6 shows M. genus, M. fermentans and M. hominis. Lane 7 shows M. genus, M. fermentans and M. penetrans. Lane 8 shows M. genus, M. hominis , and M. penetrans. Lane 9 shows M. genus, M. fermentans, M. hominins, and M. penetrans. Lane 10 is a non-CFS control sample showing no mycoplasma positive results.

This multiplex PCR was used to detect the presence of mycoplasma in patients with CFS (n=100), FMS (n=40), RA (n=60), and GWS (n=60). We found that 52, 54, 49, and 55% patients were positive for the presence of Mycoplasma genus DNA sequences respectively, while only 15% of healthy controls had a positive signal. Mycoplasma fermentans was detected in 32, 35, 23, and 36%, M. hominis was detected in 9, 8, 11, and 5%, and M. penetrans was detected in 6, 4, 7, and 3% of CFS, FMS, RA, and GWS patients respectively. M. fermentans, M. hominis, and M. penetrans were detected in 8, 3, and 2% of the healthy control samples respectively (Table 2). These results indicate that not only is mycoplasma infection occurring at a statistically significant rate in patients with CFS and related conditions over healthy controls, but also that Mycoplasma fermentans was detected at a significantly higher rate over M. hominis and M. penetrans in each sample set. M. fermentans infection averaged 32% over the combined sample sets, with the highest infection rate found in GWS patients (36%) and the lowest in RA patient samples (23%). M. hominis was detected at an average of 8% over the combined sample sets with the highest infection rate found in RA patient samples (11%) and the lowest in GWS patient samples (5%). M. penetrans was detected at an average of 5% over the combined sample sets with the highest infection rate found in RA patient samples (7%) and the lowest in GWS patient samples (3%). There were mycoplasma infections detected by the genus-specific primer set that were not identified by the three species-specific primer sets used in this assay. This indicates that patients from each sample set are infected with other mycoplasma species that remain to be identified.

Table 2. Percentages of positive results from each sample group tested.
Mycoplasma
CFS FMS RA GWS Controls
genus 52 54 49 55 15
fermentans 32 35 23 36 8
hominis 9 8 11 5 3
penetrans 6 4 7 3 2

DISCUSSION AND CONCLUSIONS

Mycoplasma detection has historically been a difficult and demanding task for many researchers. Recent advances in DNA technology have led to a wide variety of new detection techniques. One of the most important and profound contributions to molecular biology is the advent of the polymerase chain reaction (PCR). PCR has made possible in hours what only a short time ago would take weeks to achieve. Mycoplasma detection is no longer a difficult and frustrating task. PCR has given researchers and clinicians the ability to detect mycoplasma with a level of specificity and sensitivity unmatched by any other technique. Multiplex PCR is an extension of the practical use of PCR by which multiple DNA sequences may be detected in a single PCR reaction. It combines the rapid, sensitive, and specific character of conventional PCR with multiple species detection and differentiation, in effect alleviating the cost, reagent usage, and labor of individual reactions to achieve the same result. With an increased need for rapid reliable diagnostic procedures, multiplex PCR is an attractive alternative to the current detection methods available.

One application of multiplex PCR is to increase the ability of clinicians to determine the causative agent in diseases with an unknown etiology such as CFS, FMS, RA, and GWS. In a recent study, our laboratory has identified mycoplasmal infection at the genus level in 52% of CFS patients and M. fermentans infection in 34% of the same sample set (3). With the use of this multiplex PCR, the scope of our research has expanded to cover a more complete view of the role that mycoplasma infection may play in chronic illnesses with similar or overlapping symptoms. It is well documented that CFS, FMS, RA, and GWS share many common symptoms. Prior to this study there have been studies reporting on only the relationships of each of these illnesses individually. The present study suggests that mycoplasma infection may be the common link between these illnesses responsible in some degree for the similarities in symptoms experienced in each condition. Although mycoplasmas were found at a significantly higher rate in patients suffering from these illnesses over healthy controls, the role that the organisms may play in the development of symptoms remains to be determined. Using PCR based assays, M. fermentans was detected in 44.5% of saliva samples collected from healthy adults (21). This and other studies suggested that the mouth and oropharynx might be a significant reservoir of M. fermentans among human hosts, and that it may be a benign colonizer of the oral mucosa in the adult population (2). Although some believe M. fermentans to have a benign commensal relationship with the human host, other reports describe the pathogenic potential of M. fermentans (22-24). These reports found that M. fermentans was the only infectious agent that could be identified in biopsies or blood samples of previously healthy non-AIDS patients who suffered from severe chronic fatigue or a flu-like illness, and that these patients recovered from this illness after adequate antimicrobial therapy (22-24). In similar studies M. fermentans DNA was detected in 5.8% of HIV infected individuals, 15% of patients with sexually transmitted diseases, and 8% of healthy controls (25). While the prevalence of this organism in HIV infection and sexually transmitted diseases was believed to be underestimated by the investigators due to the use of long-term antimicrobial treatments, there does appear to be a correlation between the infection rate of healthy individuals found in that study and the present. It has yet to be determined why healthy individuals are infected with M. fermentans or if they will develop symptoms characteristic of CFS and related illnesses. Further studies must be undertaken to establish the relationship that mycoplasma species have with their host, and determine the possible synergism of these organisms with other biological or even chemical agents. Much remains unknown about the role of mycoplasmas in human diseases, but with further research and the advancement of molecular techniques such as multiplex PCR, we may gain more insight into the relationship between human diseases and the pathogens involved.

References

1. Grau, O., Kovacic, R., Griffais, R., et al. Development of PCR-based assays for the detection of two human mollicute species, Mycoplasma penetrans and M. hominis, Molecular and Cellular Probes 1994; 8:139-148.

2. Baseman, J.B., and Tully, J.G. Mycoplasmas; Sophisticated, reemerging, and burdened by their notoriety, Emerging Infectious Diseases 1997; 3:21-32.

3. Vojdani, A., Choppa, P.C., Tagle, C., et al. Detection of Mycoplasma genus and Mycoplasma fermentans by PCR in patients with chronic fatigue syndrome. FEMS 1998. (in press)

4. Schaeverbeke, T., Renaudin, H., Clerc, M., et al. Systematic detection of mycoplasmas by culture and polymerase chain reaction (PCR) in 209 synovial fluid samples. British Journal of Rheumatology 1997; 36: 310-314.

5. Nicolson, G.L., and Nicolson, N.L. Diagnosis and treatment of Mycoplasmal infections in Persian Gulf War Illness-CFIDS Patients. Intl Journal of Occupational Medicine, Immunology, and Toxicology 1996; 5(1): 69-78.

6. Buchwald, D. and Garrity, D. Comparison of patients with chronic fatigue syndrome, fibromyalgia, and multiple chemical sensitivities. Arch Intern Med 1994; 154: 2049-2053.

7. Ziem, G. and Done, A. Chronic fatigue, fibromyalgia, and chemical sensitivity: overlapping disorders. Arch Intern Med 1995; 154: 1913.

8. Nicolson, G.L. Chronic infections as a common etiology for many patients with chronic fatigue syndrome, fibromyalgia, and Gulf War Illness. Intern J Med 1998; 1: 42-46.

9. Kavanaugh, A.F. and Lipsky, P.E. Rheumatoid Arthritis in Clinical Immunology Principles and Practice. St. Louis:Mosby-Year Book, Inc., 1996. pp.1093-1116.

10. Kenny, G.E. and Foy, H.M. Mycoplasmal infection in Clinical Diagnosis and Management by Laboratory Methods. 18th edn. Philadelphia: W.B. Saunders 1991. pp.1264-1267.

11. Lo, S.C., Shih, J.W., Newton, P.B., et al. Virus-like infectious agent (VLIA) is a novel pathogenic Mycoplasma: Mycoplasma Incognitus. American Journal of Tropical Medicine and Hygiene 1989; 41: 586-600.

12. Hopps, H.E., Meyer, B.C., Barile, M.F., et al. Problems concerning “non-cultivable” mycoplasma contaminants in tissue cultures. Annals of the New York Academy of Science 1973; 225: 265-276.

13. Rawadi, G., and Dussurget, O. Advances in PCR-based detection of mycoplasmas contaminating cell cultures. PCR Methods and Applications 1995; 4: 199-208.

14. Razin, S. DNA probes and PCR in diagnosis of mycoplasma infections. Molecular and Cellular Probes 1994; 8: 497-511.

15. van Kuppeveld, F.J.M., van der Logt, J.T.M., Angulo, A.F., et al. Genus- and species-specific identification of Mycoplasmas by 16S rRNA amplification. Applied and Environmental Microbiology 1992; 58: 2606-2615.

16. Hopert, A., Uphoff, C.C., Wirth, M., Hauser, H., et al. Specificity and sensitivity of polymerase chain reaction (PCR) in comparison with other methods for the detection of mycoplasmas contamination in cell lines. Journal of Immunological Methods 1993; 164:91-100.

17. Wang, H., Fadl, A.A., and Khan, M.I. Multiplex PCR for avian pathogenic mycoplasmas. Molecular and Cellular Probes 1997; 11: 211-216.

18. Kulski, J.K., Khinsoe, C., Pryce, T., and Christiansen, K. Use of a multiplex PCR to detect and identify Mycobacterium avium and M. intracellulare in blood culture fluids of AIDS patients. Journal of Clinical Microbiology 1995; 33: 668-674.

19. van Kuppeveld, F.J.M., Johansson, K.E., Galama, J.M.D., et al. Detection of mycoplasma contamination in cell cultures by a Mycoplasma group-specific PCR. Applied and Environmental Microbiology 1994; 60: 149-152.

20. Hawkins, R.E., Rickman, L.S., Vermund, S.H., et al. Association of Mycoplasma and human immunodeficiency virus infection: Detection of amplified Mycoplasma fermentans DNA in blood. Journal of Infectious Diseases 1992; 165: 581-585.

21. Chingbingyong, M.I., and Hughes, C.V. Detection of Mycoplasma fermentans in human saliva with a polymerase chain reaction-based assay. Archives of Oral Biology 1996; 41: 311-314.

22. Berg, S., Luneberg, E., and Frosch, M. Development of an amplification and hybridization assay for the specific and sensitive detection of Mycoplasma fermentans. DNA Molecular and Cellular Probes 1996; 10: 7-14.

23. Lo, S.C., Buchholz, C.L., Wear, D.J., et al. Histopathology and doxycycline treatment in a previously healthy non-AIDS patient systematically infected by Mycoplasma fermentans (incognitus strain). Modern Pathology 1991; 4: 750-754.

24. Beecham, H.J., Lo, S.C., Lewis, D.E., et al. Recovery from fulminant infection with Mycoplasma fermentans (incognitus strain) in non-immunocompromised host. Lancet 1991; 338: 1014-1015.

25. Kovacic, R., Launay, V., Tuppin, P., et al. Search for the presence of six mycoplasma species in peripheral blood mononuclear cells of subjects seropositive and seronegative for human immunodeficiency virus. Journal of Clinical Microbiology 1996; 34:1808-1810.