- Email: [email protected]
Case of Mycobacterium tuberculosis meningitis: Gram staining as a useful initial diagnostic clue for tuberculous meningitis
J Infect Chemother (2012) 18:931–936 DOI 10.1007/s10156-012-0382-y
CASE REPORT
Case of Mycobacterium tuberculosis meningitis: Gram staining as a useful initial diagnostic clue for tuberculous meningitis Sayoko Kawakami • Yasuyosi Kawamura • Kyouhei Nishiyama • Hiroki Hatanaka Ryuichi Fujisaki • Yasuo Ono • Yukihisa Miyazawa • Hajime Nishiya
•
Received: 26 May 2011 / Accepted: 25 January 2012 / Published online: 5 April 2012 Ó Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases 2012
Abstract A 32-year-old man was admitted to our hospital because of fever, headache, and loss of consciousness. Four days before admission, he had had difficulty speaking. On the day of admission, his colleague had found him to be unconscious and lying on his back. He was admitted to our hospital. The temperature at the eardrum was 35.2°C. Neurologic evaluation was negative. Computed tomography (CT) scan of the brain showed slight ventricular enlargement bilaterally. An X-ray film of the chest showed no abnormality. On the second hospital day, neck stiffness was noted. The cerebrospinal fluid (CSF) contained 870 white cells/ll, most of which were neutrophils; the glucose level in the CSF was 10 mg/dl, and the protein level was 140 mg/dl. Stained smears of the CSF, including Gram staining and India-ink preparations, disclosed no microorganisms. Capsular antigen tests for several bacteria were negative. Antimicrobial agents were started. However, by
changing the microscope focus slightly while viewing Gram stains of the CSF, we could see brightened and Gram-positive bacilli that had been phagocytosed by neutrophils. This finding suggested the presence of Mycobacterium tuberculosis. Ziehl–Neelsen staining of the CSF and gastric juice revealed anti-acid bacilli. Polymerase chain reaction for M. tuberculosis in the gastric juice was positive. This case showed that Gram staining could be useful as an initial adjunct for the diagnosis of tuberculous meningitis, particularly when the CSF shows predominantly neutrocytic pleocytosis, but no other evidence of bacterial meningitis. Keywords Gram stain Mycobacterium tuberculosis Tuberculous meningitis
Introduction S. Kawakami Y. Miyazawa Department of Central Clinical Laboratory, Teikyo University Hospital, Tokyo, Japan Y. Kawamura K. Nishiyama H. Hatanaka Department of Neurology, Teikyo University Hospital, Tokyo, Japan R. Fujisaki H. Nishiya Department of Internal Medicine, Teikyo University Hospital, Tokyo, Japan Y. Ono Department of Bacteriology, Teikyo University Hospital, Tokyo, Japan H. Nishiya (&) Department of Internal Medicine, Teikyo University School of Medicine, Kaga 2-11-2, Itabashi-ku, Tokyo 133, Japan e-mail: [email protected]
Establishing a diagnosis of intracranial tuberculosis, i.e., tuberculous meningitis or tuberculoma, is notoriously difficult. Prompt diagnosis and treatment of tuberculous meningitis saves lives. However, about 30% of patients with tuberculous meningitis die despite receiving antituberculosis chemotherapy [1]. Delays in diagnosis and treatment are regarded as major contributing factors to the high mortality reported in many recent series [2]. We have experienced cases in which tuberculosis was first suggested by Gram staining of sputum or pus, and we have reported that Gram staining is an easy to use and rapid initial diagnostic tool with which to recognize highly infective pulmonary and dermal tuberculosis [3]. We report here a patient in whom Gram staining of the cerebrospinal fluid (CSF) presented a useful initial diagnostic clue for Mycobacterium tuberculosis infection, which was confirmed by
123
932
Ziehl–Neelsen staining and diagnosed by polymerase chain reaction (PCR) of the gastric fluid. To our knowledge, this is the first report of tuberculous meningitis in which the diagnosis had been predicted by Gram staining.
Case report A 32-year-old man was admitted to our hospital because of fever, headache, and loss of consciousness. He had been well until two and a half weeks previously, when he had experienced cough and vomiting. He stayed home from work. Ten days before admission to our hospital, he had visited another hospital and was diagnosed as having enterocolitis. Four days before admission, he had had difficulty speaking over the phone and his supervisor at work told him to come to work. After he went back to work, the patient was talking to himself and his eye movements were not focused, as reported by his colleague. On the day before admission, he did not go to work and his colleague could not reach him by telephone. On the next day, his colleague went to his apartment and found him to be unconsciousness lying on his back. He was admitted to our hospital. His colleague reported his smoking history to be a half pack of cigarettes daily for many years. The temperature at his eardrum was 35.2°C. His pulse was 111 beats/min, and respiratory rate 36/min. His blood pressure was 100/50 mmHg. On examination, the patient was somnolent. His eyes showed prompt light reflex. No jaundice or spider angiomas were noted. No lymphadenopathy was found. There was no abnormality in the lungs, heart, or abdomen. His bilateral upper limbs showed abnormal movements similar to flapping tremor. Neurologic evaluation was negative. An electrocardiogram showed a normal rhythm at a rate of 111 beats/min. Urinalysis was normal. Laboratory tests were performed (Table 1). Erythrocytosis was found. The white cell count was elevated with an increased percentage of neutrophils. C-reactive protein (CRP) was slightly elevated. Renal failure and slight liver dysfunction were noted. The serum NH3 and creatine kinase levels were increased. Tests for syphilis on admission were positive. However, the levels on the serologic test for syphilis (STS) and Treponema pallidum hemagglutination assay (TPHA) on the next day revealed the existence of a prior infection. A computed tomography (CT) scan of the brain at the time of admission showed slight ventricular enlargement bilaterally. An X-ray film of the chest showed no abnormality. On the second hospital day, neck stiffness was noted in the morning and Kernig’s sign was positive. Lumbar puncture was performed; the fluid contained 870 white cells/ll, most of which were neutrophils; the CSF showed a
123
J Infect Chemother (2012) 18:931–936
low glucose level and a high protein level. Gram-stained smears of the CSF and an India-ink preparation disclosed no microorganisms. Urine pneumococcal antigen was negative. Capsular antigen tests of the CSF for Streptococcus pneumoniae, Group B Streptococcus, Haemophilus influenzae type b and Neisseria meningitidis were negative. However, neutrocytic pleocytosis and the low glucose level in the CSF suggested bacterial meningitis, and antimicrobial agents (meropenem and vancomycin) were started. The Gram staining of the CSF was rechecked by a skilled technician who had experience in finding tubercle bacilli in Gram-stained sputa. By changing the microscope focus slightly while observing the slides, brightened rods or colorless bacilli (slightly shorter focus) and Gram-positive cord-like bacilli (slightly longer focus), which had been phagocytosed by neutrophils, were seen (Fig. 1). This finding suggested the presence of M. tuberculosis. Ziehl–Neelsen staining of the CSF revealed anti-acid bacilli, most of which had been phagocytosed by neutrophils (Fig. 2). Ziehl–Neelsen staining of gastric juice also showed anti-acid bacilli. Polymerase chain reaction (PCR) of the gastric juice was positive for tubercle bacillus, but that of the CSF was negative. The level of adenosine deaminase (ADA) in the CSF was 21.5 IU/l. The antitubercular agents isoniazid (INH), rifampicin (RFP), pyrazinamide, and ethambuto, were started. On the third hospital day, his consciousness got a little better. His temperature was 37.5°C. The white blood cell count and renal function showed normal levels. The Ziehl– Neelsen staining of sputum expectorated on that day showed no tubercle bacilli. The patient was transferred to a hospital specializing in tuberculosis. Subsequently, it was proven that culture of the CSF showed no bacterial growth, while culture of the CSF, gastric juice, and sputa later revealed M. tuberculosis growth. After transfer to the specialist hospital, he was administered RFP (450 mg/day), INH (300 mg/day), and prednisolone (15 mg/day), and his condition improved. We were informed that the total schedule of RFP and INH administration was for 1 year.
Discussion Establishing a diagnosis of tuberculous meningitis is difficult. The clinical diagnosis begins with a high index of suspicion in patients with a compatible clinical presentation and suspicious laboratory findings, and the proper examination and culture of the CSF is the key to diagnosis in the great majority of cases [4]. The cornerstone of diagnosis is examination of the CSF. The classic CSF abnormalities in tuberculous meningitis are as follows: (1) elevated opening pressure, (2) lymphocytic pleocytosis
J Infect Chemother (2012) 18:931–936
933
Table 1 Laboratory values 2011/03/16
2011/03/17
2011/03/18
RBC
690 9 104/ll
Hb
20.8 g/dl
Seg
96%
Hb
14.6 g/dl
Ht
57.5%
Ly
1%
Ht
43.9%
MCV
83.3 fl
Mono
3%
MCV
89.0 fl
MCH
30.1 pg
Eos
0%
MCHC
33.3%
MCHC
36.2%
Bas
0%
WBC
7900/ll
Plt
24.2 9 104/ll
CRP
1.86 mg/dl
Stab
3%
WBC
18900/ll
RBC
493 9 104/ll
WBC
21100/ll
BUN
50.3 mg/dl
Seg
90%
PT
13.9 s
Cr
0.94 mg/dl
Ly
5%
PT%
88.0%
Endotoxin
3.0 pg/ml
Mono
2%
APTT TP
23.5 s 7.1 g/dl
b-D glucan CSF
6 pg/ml
Eos Bas
0% 0%
T-Bil
0.80 mg/dl
WBC
870/ll
BUN
16.1 mg/dl
D-Bi1
0.14 mg/dl
Mononuclear cells [ neutrophils
Cr
0.50 mg/dl
AST
33 IU/l
Glucose
10 mg/dl
CRP
0.98 mg/dl
ALT
42 IU/l
Protein
140 mg/dl
TSH
0.076 lU/ml
LDH
372 IU/l
Chloride
108 mEq/l
F-T4
1.0 ng/dl
ALP
325 IU/l
Adenosine deaminase
21.5 IU/l
HBsAb
0.5 mIU/ml
c-GTP
28 U/l
STS
Onefold
CK
680 U/l
TPHA
640-fold
BUN
93.9 mg/dl
HIV 1-RNA
Negative
Cr
2.42 mg/dl
Na
130 mEq/l
K
4.8 mEq/l
Cl
84 mEq/l
Ca
4.3 mEq/l
CRP BS
2.43 mg/dl 150 mg/dl
Amy
168 U/l
NH3
89 lg/dl
Vit B12
950 pg/ml
Vit B1
37 ng/ml
HIV-Ab
(-)
RPR
(?)
TPHA
(?)
CSF cerebrospinal fluid, STS serologic test for syphilis, TPHA Treponema pallidum hemagglutination assay, RBC red blood cell, Hb hemoglobin, Ht hematocrit, Plt platelets, WBC white blood cell, Stab stab cell, Seg segmented cell, Lym lymphocyte, Mono moncyte, Eos eosinophil, Bso basophil, PT prothrombin time, APTT activated partial thromboplastin time, TP total protein, T-Bil total bilirubin, D-Bil direct bilirubin, AST aspartate aminotransferase, ALT alanine aminotransaminase, LDH lactate dehydrogenase, ALP alkaline phosphatase, gamma-GTP gammaglutamyl transpeptidase, CK creatine kinase, BUN blood urea nitrogen, Cr creatinine, BS blood sugar, Amy amylase, NH3 ammonia, RPR rapid plasma regain card test, TSH thyroid stimulating hormone, F-T4 free thyroxin, HBs-Ab hepatitis B surface antibody
(10–500 cells/ll), (3) elevated protein concentration in the range of 1–5 g/l (100–500 mg/dl), and (4) reduced glucose concentration in the range of 1.1–2.2 mmol/l (20–40 mg/ dl). The combination of unrelenting headache, stiff neck, fatigue, night sweats, and fever with CSF lymphocytic pleocytosis and a mildly reduced glucose concentration is
highly suspicious for tuberculous meningitis [5]. Culture remains the ‘‘gold standard’’ in making the diagnosis of tuberculous meningitis. However, it takes 4–8 weeks to perform cultures of the CSF in order to identify the organism, and the culture is positive in only about 50% of adults with tuberculous meningitis [5] and in 38–88% of
123
934
Fig. 1 Gram staining of the cerebrospinal fluid (91000) Mycobacterium tuberculosis organisms that had been phagocytosed by neutrophils are shown in pairs and are stained colorless (on the left side) or Gram-positive (on the right side). Changing the microscope
Fig. 2 Ziehl–Neelsen staining of the cerebrospinal fluid (91000) Mycobacterium tuberculosis organisms stained in pink had been phagocytosed by neutrophils
overall patients with tuberculous meningitis [6]. Smears of CSF contain acid-fast bacilli in 10–22% of patients with tuberculous meningitis [6] and stains of sediment reveal acid-fast bacilli in 37% of cases on initial examination [7]. These low yields of culture and smear positivity for M. tuberculosis make the diagnosis of tuberculous meningitis difficult.
123
J Infect Chemother (2012) 18:931–936
focus slightly while observing the slides revealed brightened or colorless bacilli (slightly shorter focus; left sides) and Gram-positive bacilli (slightly longer focus; right sides)
Various antibody and antigen detection assays using CSF are available to make the presumptive diagnosis of tuberculous meningitis: currently, detection of mycobacterial DNA by PCR has become standard and is routinely performed, often being the first confirmatory test. The detection of M. tuberculosis DNA by PCR has a sensitivity of 70–80% [4], but rates of false-positivity reach 10% [8]. In our patient, the PCR was positive for the gastric juice and negative for the CSF. The exact reason why the PCR was negative for the CSF is not clear, but this finding may be related to the relatively low sensitivity of PCR, with 20% to 30% of cases of tuberculosis meningitis being PCR-negative. In our experience, the PCR for tubercle bacilli in the CSF is less sensitive than that in sputa, which suggests the possible existence of some blocking agents in the CSF against the PCR test. The low PCR sensitivity for CSF in the present patient might have been related to the fact that most of the tubercle bacilli in the CSF were phagocytosed by neutrophils. A PCR test for the sputa was not done. But tubercle bacilli were cultured in the sputa 2 weeks after the patient’s admission to our hospital. The patient’s chest X-ray showed no abnormal findings. After he was transferred to the specialist tuberculosis hospital, we were informed that the culture of his sputa was positive for tubercle bacilli. These findings indicated that he had a small lesion of pulmonary tuberculosis in his lung which our chest X-ray did not show. Adenosine deaminase (ADA) is released by T lymphocytes, and an increased ADA level in the CSF is found in
J Infect Chemother (2012) 18:931–936
tuberculous meningitis, but not in viral or bacterial meningitis. Determination of the ADA concentration may be a sensitive test, but it has low specificity [9]. In our patient, the ADA level in the CSF was increased. However, it took some time to get the data, which decreased its usefulness. Gram staining is a useful technique for detecting bacteria, but it is highly questionable whether Gram staining can detect M. tuberculosis. The detection of this organism generally requires special staining, such as Ziehl–Neelsen staining. There have been no descriptions of the usefulness of Gram staining in diagnosing tuberculous meningitis [5]. Major textbooks of microbiology, clinical pulmonary medicine, and clinical pathology generally fail to describe in detail the staining characteristics of M. tuberculosis. They tend to describe the results of acid-fast techniques, either ignoring the Gram-stain appearance of the tubercle bacillus or reporting a faint Gram-positive quality [10]. Hinson et al. [10] and Trifiro et al. [11] observed tubercle bacilli as ghost mycobacteria in Gram staining of clinical samples. We have experienced cases in which tuberculosis was first suggested by Gram staining of sputum or pus and confirmed by Ziehl–Neelsen staining, and we have reported that Gram staining is an easy to use and rapid initial diagnostic tool with which to recognize highly infective pulmonary and dermal tuberculosis [3]. To detect colorless tubercle bacilli in clinical samples, changing the focus of the microscope while observing the slides is indispensable [3]. The ease of the procedure was confirmed by medical students who were not skilled in the technique of Gram staining. They could easily find tubercle bacilli after they were informed about the procedure. We have recently shown diagrams of the procedures [12]. In tuberculous meningitis, lymphocytic predominance in the CSF is usually observed, although it should be noted that neutrophils may predominate, particularly during the first 10 days of symptoms [6]. One-quarter of cases demonstrate polymorphonuclear pleocytosis, usually early in the course. These situations raise diagnostic uncertainty. Thwaites et al. [1] showed that simple clinical and laboratory data could help in diagnosing tuberculous meningitis in adults. They presented five features predictive of a diagnosis of tuberculous meningitis: age, length of history, WBC count, total CSF white-cell count, and CSF neutrophil proportion. The length of the history in our patient was uncertain and the blood WBC count and percentage of neutrophils in the CSF suggested bacterial meningitis. Hinson et al. [10] reported that purulent sputum without organisms on Gram stain should warrant a search for acidfast bacilli. Our laboratory technologist had suspected the presence of M. tuberculosis in the CSF which showed polymorphonuclear pleocytosis, no bacteria on Gram staining, and negative capsular antigen test for bacteria. The technologist reevaluated the Gram-stained smears by
935
changing the focus of the microscope and found colorless and Gram-positive tubercle bacilli that had been phagocytosed by neutrophils. Ziehl–Neelsen staining, performed after the Gram staining, showed M. tuberculosis in the CSF (Fig. 2). The majority of M. tuberculosis organisms in the CSF were found inside neutrophils on Gram staining. The number of tubercle bacilli-phagocytic neutrophils was small (tubercle bacilli-phagocytic neutrophils/neutrophils/ 100 WBC in the CSF; 2/92/100) on Gram stain, and was one-half of that seen on Ziehl–Neelsen staining (data not shown). The small percentage of tubercle bacilli-phagocytic neutrophils indicates that careful attention must be paid to find bacilli in Gram-stained specimens. Our case showed for the first time that Gram staining is a useful tool to predict the presence of tuberculous meningitis. However, one possible limitation is that its usefulness may be restricted to cases in which the CSF shows predominantly neutrocytic pleocytosis but no other evidence of bacterial meningitis. Further studies are needed to clarify the usefulness of Gram staining in which the CSF shows lymphocytic pleocytosis. The ease of the procedure of the focus-changing technique when viewing Gram-stained specimens will contribute to its use as an initial diagnostic tool before confirming the diagnosis of tubercular meningitis with predominantly neutrocytic pleocytosis. This procedure would be especially useful in developing countries.
References 1. Thwaites GE, Chau TT, Mai NT, et al. Tuberculous meningitis. J Neurol Neurosurg Psychiatry. 2000;68:289–99. 2. Verdon R, Chevret S, Laissy JP, Wolff M. Tuberculous meningitis in adults: review of 48 cases. Clin Infect Dis. 1996;22: 982–8. 3. Atsukawa Y, Kawakami S, Asahara M, Ishigaki S, TanakaT, Ono Y, et al. The usefulness of focus changing examination in Gram staining as initial diagnostic clue for infective tuberculosis. J Infect Chemother. 2011;17:571–74. 4. Leonared JM, Des Prez RM. Tuberculous meningitis. Infect Dis Clin North Am. 1990;4:769–87. 5. Raviglione MC, O’Brien RJ. Tuberculosis (chapter 158) In: Harrison’s principles of internal medicine. 17th ed. New York: The McGraw-Hill Companies, Inc.; 2008. p. 2634. 6. Behlau I, Ellner JJ. Chronic meningitis (chapter 82) In: Principles and practice of infectious diseases. 6th ed. Philadelphia: Elsevier Churchill Livingstone; 2005. p. 1136. 7. Fitzgerald D, Haas D. Mycobacterium tuberculosis (chapter 248) In: Principles and practice of infectious diseases. 6th ed. Philadelphia: Elsevier Churchill Livingstone; 2005. p. 2877. 8. Raviglione MC, O’Brien RJ. Tuberculosis (chapter 158) In: Harrison’s principles of internal medicine. 17th ed. New York: The McGraw-Hill Companies, Inc.; 2008. p. 1012. 9. Choi SH, Kim YS, Bae IG, Chung JW. The possible role of cerebrospinal fluid adenosine deaminase activity in the diagnosis
123
936 of tuberculous meningitis in adults. Clin Neurol Neurosurg. 2002;104:10–5. 10. Hinson JM, Bradsher RW, Bodner SJ. Gram-stain neutrality of Mycobacterium tuberculosis. Am Rev Respir Dis. 1981;123: 365–6. 11. Trifiro S, Bourgault AM, Lebel F. Ghost mycobacteria on Gram stain. J Clin Microbiol. 1990;28:146–7.
123
J Infect Chemother (2012) 18:931–936 12. Atsukawa Y, Kawakami S, Ono Y, Fujisaki R, Miyazawa Y, Nishiya H. Three cases of Mycobacterium tuberculosis infection initially recognized by focus changing examination in gram staining. In: Roy PK, editor. Insight and control of infectious disease in global scenario. InTech—Open Access Company (ISBN 978-953-51-0319-6) (in press).
CASE REPORT
Case of Mycobacterium tuberculosis meningitis: Gram staining as a useful initial diagnostic clue for tuberculous meningitis Sayoko Kawakami • Yasuyosi Kawamura • Kyouhei Nishiyama • Hiroki Hatanaka Ryuichi Fujisaki • Yasuo Ono • Yukihisa Miyazawa • Hajime Nishiya
•
Received: 26 May 2011 / Accepted: 25 January 2012 / Published online: 5 April 2012 Ó Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases 2012
Abstract A 32-year-old man was admitted to our hospital because of fever, headache, and loss of consciousness. Four days before admission, he had had difficulty speaking. On the day of admission, his colleague had found him to be unconscious and lying on his back. He was admitted to our hospital. The temperature at the eardrum was 35.2°C. Neurologic evaluation was negative. Computed tomography (CT) scan of the brain showed slight ventricular enlargement bilaterally. An X-ray film of the chest showed no abnormality. On the second hospital day, neck stiffness was noted. The cerebrospinal fluid (CSF) contained 870 white cells/ll, most of which were neutrophils; the glucose level in the CSF was 10 mg/dl, and the protein level was 140 mg/dl. Stained smears of the CSF, including Gram staining and India-ink preparations, disclosed no microorganisms. Capsular antigen tests for several bacteria were negative. Antimicrobial agents were started. However, by
changing the microscope focus slightly while viewing Gram stains of the CSF, we could see brightened and Gram-positive bacilli that had been phagocytosed by neutrophils. This finding suggested the presence of Mycobacterium tuberculosis. Ziehl–Neelsen staining of the CSF and gastric juice revealed anti-acid bacilli. Polymerase chain reaction for M. tuberculosis in the gastric juice was positive. This case showed that Gram staining could be useful as an initial adjunct for the diagnosis of tuberculous meningitis, particularly when the CSF shows predominantly neutrocytic pleocytosis, but no other evidence of bacterial meningitis. Keywords Gram stain Mycobacterium tuberculosis Tuberculous meningitis
Introduction S. Kawakami Y. Miyazawa Department of Central Clinical Laboratory, Teikyo University Hospital, Tokyo, Japan Y. Kawamura K. Nishiyama H. Hatanaka Department of Neurology, Teikyo University Hospital, Tokyo, Japan R. Fujisaki H. Nishiya Department of Internal Medicine, Teikyo University Hospital, Tokyo, Japan Y. Ono Department of Bacteriology, Teikyo University Hospital, Tokyo, Japan H. Nishiya (&) Department of Internal Medicine, Teikyo University School of Medicine, Kaga 2-11-2, Itabashi-ku, Tokyo 133, Japan e-mail: [email protected]
Establishing a diagnosis of intracranial tuberculosis, i.e., tuberculous meningitis or tuberculoma, is notoriously difficult. Prompt diagnosis and treatment of tuberculous meningitis saves lives. However, about 30% of patients with tuberculous meningitis die despite receiving antituberculosis chemotherapy [1]. Delays in diagnosis and treatment are regarded as major contributing factors to the high mortality reported in many recent series [2]. We have experienced cases in which tuberculosis was first suggested by Gram staining of sputum or pus, and we have reported that Gram staining is an easy to use and rapid initial diagnostic tool with which to recognize highly infective pulmonary and dermal tuberculosis [3]. We report here a patient in whom Gram staining of the cerebrospinal fluid (CSF) presented a useful initial diagnostic clue for Mycobacterium tuberculosis infection, which was confirmed by
123
932
Ziehl–Neelsen staining and diagnosed by polymerase chain reaction (PCR) of the gastric fluid. To our knowledge, this is the first report of tuberculous meningitis in which the diagnosis had been predicted by Gram staining.
Case report A 32-year-old man was admitted to our hospital because of fever, headache, and loss of consciousness. He had been well until two and a half weeks previously, when he had experienced cough and vomiting. He stayed home from work. Ten days before admission to our hospital, he had visited another hospital and was diagnosed as having enterocolitis. Four days before admission, he had had difficulty speaking over the phone and his supervisor at work told him to come to work. After he went back to work, the patient was talking to himself and his eye movements were not focused, as reported by his colleague. On the day before admission, he did not go to work and his colleague could not reach him by telephone. On the next day, his colleague went to his apartment and found him to be unconsciousness lying on his back. He was admitted to our hospital. His colleague reported his smoking history to be a half pack of cigarettes daily for many years. The temperature at his eardrum was 35.2°C. His pulse was 111 beats/min, and respiratory rate 36/min. His blood pressure was 100/50 mmHg. On examination, the patient was somnolent. His eyes showed prompt light reflex. No jaundice or spider angiomas were noted. No lymphadenopathy was found. There was no abnormality in the lungs, heart, or abdomen. His bilateral upper limbs showed abnormal movements similar to flapping tremor. Neurologic evaluation was negative. An electrocardiogram showed a normal rhythm at a rate of 111 beats/min. Urinalysis was normal. Laboratory tests were performed (Table 1). Erythrocytosis was found. The white cell count was elevated with an increased percentage of neutrophils. C-reactive protein (CRP) was slightly elevated. Renal failure and slight liver dysfunction were noted. The serum NH3 and creatine kinase levels were increased. Tests for syphilis on admission were positive. However, the levels on the serologic test for syphilis (STS) and Treponema pallidum hemagglutination assay (TPHA) on the next day revealed the existence of a prior infection. A computed tomography (CT) scan of the brain at the time of admission showed slight ventricular enlargement bilaterally. An X-ray film of the chest showed no abnormality. On the second hospital day, neck stiffness was noted in the morning and Kernig’s sign was positive. Lumbar puncture was performed; the fluid contained 870 white cells/ll, most of which were neutrophils; the CSF showed a
123
J Infect Chemother (2012) 18:931–936
low glucose level and a high protein level. Gram-stained smears of the CSF and an India-ink preparation disclosed no microorganisms. Urine pneumococcal antigen was negative. Capsular antigen tests of the CSF for Streptococcus pneumoniae, Group B Streptococcus, Haemophilus influenzae type b and Neisseria meningitidis were negative. However, neutrocytic pleocytosis and the low glucose level in the CSF suggested bacterial meningitis, and antimicrobial agents (meropenem and vancomycin) were started. The Gram staining of the CSF was rechecked by a skilled technician who had experience in finding tubercle bacilli in Gram-stained sputa. By changing the microscope focus slightly while observing the slides, brightened rods or colorless bacilli (slightly shorter focus) and Gram-positive cord-like bacilli (slightly longer focus), which had been phagocytosed by neutrophils, were seen (Fig. 1). This finding suggested the presence of M. tuberculosis. Ziehl–Neelsen staining of the CSF revealed anti-acid bacilli, most of which had been phagocytosed by neutrophils (Fig. 2). Ziehl–Neelsen staining of gastric juice also showed anti-acid bacilli. Polymerase chain reaction (PCR) of the gastric juice was positive for tubercle bacillus, but that of the CSF was negative. The level of adenosine deaminase (ADA) in the CSF was 21.5 IU/l. The antitubercular agents isoniazid (INH), rifampicin (RFP), pyrazinamide, and ethambuto, were started. On the third hospital day, his consciousness got a little better. His temperature was 37.5°C. The white blood cell count and renal function showed normal levels. The Ziehl– Neelsen staining of sputum expectorated on that day showed no tubercle bacilli. The patient was transferred to a hospital specializing in tuberculosis. Subsequently, it was proven that culture of the CSF showed no bacterial growth, while culture of the CSF, gastric juice, and sputa later revealed M. tuberculosis growth. After transfer to the specialist hospital, he was administered RFP (450 mg/day), INH (300 mg/day), and prednisolone (15 mg/day), and his condition improved. We were informed that the total schedule of RFP and INH administration was for 1 year.
Discussion Establishing a diagnosis of tuberculous meningitis is difficult. The clinical diagnosis begins with a high index of suspicion in patients with a compatible clinical presentation and suspicious laboratory findings, and the proper examination and culture of the CSF is the key to diagnosis in the great majority of cases [4]. The cornerstone of diagnosis is examination of the CSF. The classic CSF abnormalities in tuberculous meningitis are as follows: (1) elevated opening pressure, (2) lymphocytic pleocytosis
J Infect Chemother (2012) 18:931–936
933
Table 1 Laboratory values 2011/03/16
2011/03/17
2011/03/18
RBC
690 9 104/ll
Hb
20.8 g/dl
Seg
96%
Hb
14.6 g/dl
Ht
57.5%
Ly
1%
Ht
43.9%
MCV
83.3 fl
Mono
3%
MCV
89.0 fl
MCH
30.1 pg
Eos
0%
MCHC
33.3%
MCHC
36.2%
Bas
0%
WBC
7900/ll
Plt
24.2 9 104/ll
CRP
1.86 mg/dl
Stab
3%
WBC
18900/ll
RBC
493 9 104/ll
WBC
21100/ll
BUN
50.3 mg/dl
Seg
90%
PT
13.9 s
Cr
0.94 mg/dl
Ly
5%
PT%
88.0%
Endotoxin
3.0 pg/ml
Mono
2%
APTT TP
23.5 s 7.1 g/dl
b-D glucan CSF
6 pg/ml
Eos Bas
0% 0%
T-Bil
0.80 mg/dl
WBC
870/ll
BUN
16.1 mg/dl
D-Bi1
0.14 mg/dl
Mononuclear cells [ neutrophils
Cr
0.50 mg/dl
AST
33 IU/l
Glucose
10 mg/dl
CRP
0.98 mg/dl
ALT
42 IU/l
Protein
140 mg/dl
TSH
0.076 lU/ml
LDH
372 IU/l
Chloride
108 mEq/l
F-T4
1.0 ng/dl
ALP
325 IU/l
Adenosine deaminase
21.5 IU/l
HBsAb
0.5 mIU/ml
c-GTP
28 U/l
STS
Onefold
CK
680 U/l
TPHA
640-fold
BUN
93.9 mg/dl
HIV 1-RNA
Negative
Cr
2.42 mg/dl
Na
130 mEq/l
K
4.8 mEq/l
Cl
84 mEq/l
Ca
4.3 mEq/l
CRP BS
2.43 mg/dl 150 mg/dl
Amy
168 U/l
NH3
89 lg/dl
Vit B12
950 pg/ml
Vit B1
37 ng/ml
HIV-Ab
(-)
RPR
(?)
TPHA
(?)
CSF cerebrospinal fluid, STS serologic test for syphilis, TPHA Treponema pallidum hemagglutination assay, RBC red blood cell, Hb hemoglobin, Ht hematocrit, Plt platelets, WBC white blood cell, Stab stab cell, Seg segmented cell, Lym lymphocyte, Mono moncyte, Eos eosinophil, Bso basophil, PT prothrombin time, APTT activated partial thromboplastin time, TP total protein, T-Bil total bilirubin, D-Bil direct bilirubin, AST aspartate aminotransferase, ALT alanine aminotransaminase, LDH lactate dehydrogenase, ALP alkaline phosphatase, gamma-GTP gammaglutamyl transpeptidase, CK creatine kinase, BUN blood urea nitrogen, Cr creatinine, BS blood sugar, Amy amylase, NH3 ammonia, RPR rapid plasma regain card test, TSH thyroid stimulating hormone, F-T4 free thyroxin, HBs-Ab hepatitis B surface antibody
(10–500 cells/ll), (3) elevated protein concentration in the range of 1–5 g/l (100–500 mg/dl), and (4) reduced glucose concentration in the range of 1.1–2.2 mmol/l (20–40 mg/ dl). The combination of unrelenting headache, stiff neck, fatigue, night sweats, and fever with CSF lymphocytic pleocytosis and a mildly reduced glucose concentration is
highly suspicious for tuberculous meningitis [5]. Culture remains the ‘‘gold standard’’ in making the diagnosis of tuberculous meningitis. However, it takes 4–8 weeks to perform cultures of the CSF in order to identify the organism, and the culture is positive in only about 50% of adults with tuberculous meningitis [5] and in 38–88% of
123
934
Fig. 1 Gram staining of the cerebrospinal fluid (91000) Mycobacterium tuberculosis organisms that had been phagocytosed by neutrophils are shown in pairs and are stained colorless (on the left side) or Gram-positive (on the right side). Changing the microscope
Fig. 2 Ziehl–Neelsen staining of the cerebrospinal fluid (91000) Mycobacterium tuberculosis organisms stained in pink had been phagocytosed by neutrophils
overall patients with tuberculous meningitis [6]. Smears of CSF contain acid-fast bacilli in 10–22% of patients with tuberculous meningitis [6] and stains of sediment reveal acid-fast bacilli in 37% of cases on initial examination [7]. These low yields of culture and smear positivity for M. tuberculosis make the diagnosis of tuberculous meningitis difficult.
123
J Infect Chemother (2012) 18:931–936
focus slightly while observing the slides revealed brightened or colorless bacilli (slightly shorter focus; left sides) and Gram-positive bacilli (slightly longer focus; right sides)
Various antibody and antigen detection assays using CSF are available to make the presumptive diagnosis of tuberculous meningitis: currently, detection of mycobacterial DNA by PCR has become standard and is routinely performed, often being the first confirmatory test. The detection of M. tuberculosis DNA by PCR has a sensitivity of 70–80% [4], but rates of false-positivity reach 10% [8]. In our patient, the PCR was positive for the gastric juice and negative for the CSF. The exact reason why the PCR was negative for the CSF is not clear, but this finding may be related to the relatively low sensitivity of PCR, with 20% to 30% of cases of tuberculosis meningitis being PCR-negative. In our experience, the PCR for tubercle bacilli in the CSF is less sensitive than that in sputa, which suggests the possible existence of some blocking agents in the CSF against the PCR test. The low PCR sensitivity for CSF in the present patient might have been related to the fact that most of the tubercle bacilli in the CSF were phagocytosed by neutrophils. A PCR test for the sputa was not done. But tubercle bacilli were cultured in the sputa 2 weeks after the patient’s admission to our hospital. The patient’s chest X-ray showed no abnormal findings. After he was transferred to the specialist tuberculosis hospital, we were informed that the culture of his sputa was positive for tubercle bacilli. These findings indicated that he had a small lesion of pulmonary tuberculosis in his lung which our chest X-ray did not show. Adenosine deaminase (ADA) is released by T lymphocytes, and an increased ADA level in the CSF is found in
J Infect Chemother (2012) 18:931–936
tuberculous meningitis, but not in viral or bacterial meningitis. Determination of the ADA concentration may be a sensitive test, but it has low specificity [9]. In our patient, the ADA level in the CSF was increased. However, it took some time to get the data, which decreased its usefulness. Gram staining is a useful technique for detecting bacteria, but it is highly questionable whether Gram staining can detect M. tuberculosis. The detection of this organism generally requires special staining, such as Ziehl–Neelsen staining. There have been no descriptions of the usefulness of Gram staining in diagnosing tuberculous meningitis [5]. Major textbooks of microbiology, clinical pulmonary medicine, and clinical pathology generally fail to describe in detail the staining characteristics of M. tuberculosis. They tend to describe the results of acid-fast techniques, either ignoring the Gram-stain appearance of the tubercle bacillus or reporting a faint Gram-positive quality [10]. Hinson et al. [10] and Trifiro et al. [11] observed tubercle bacilli as ghost mycobacteria in Gram staining of clinical samples. We have experienced cases in which tuberculosis was first suggested by Gram staining of sputum or pus and confirmed by Ziehl–Neelsen staining, and we have reported that Gram staining is an easy to use and rapid initial diagnostic tool with which to recognize highly infective pulmonary and dermal tuberculosis [3]. To detect colorless tubercle bacilli in clinical samples, changing the focus of the microscope while observing the slides is indispensable [3]. The ease of the procedure was confirmed by medical students who were not skilled in the technique of Gram staining. They could easily find tubercle bacilli after they were informed about the procedure. We have recently shown diagrams of the procedures [12]. In tuberculous meningitis, lymphocytic predominance in the CSF is usually observed, although it should be noted that neutrophils may predominate, particularly during the first 10 days of symptoms [6]. One-quarter of cases demonstrate polymorphonuclear pleocytosis, usually early in the course. These situations raise diagnostic uncertainty. Thwaites et al. [1] showed that simple clinical and laboratory data could help in diagnosing tuberculous meningitis in adults. They presented five features predictive of a diagnosis of tuberculous meningitis: age, length of history, WBC count, total CSF white-cell count, and CSF neutrophil proportion. The length of the history in our patient was uncertain and the blood WBC count and percentage of neutrophils in the CSF suggested bacterial meningitis. Hinson et al. [10] reported that purulent sputum without organisms on Gram stain should warrant a search for acidfast bacilli. Our laboratory technologist had suspected the presence of M. tuberculosis in the CSF which showed polymorphonuclear pleocytosis, no bacteria on Gram staining, and negative capsular antigen test for bacteria. The technologist reevaluated the Gram-stained smears by
935
changing the focus of the microscope and found colorless and Gram-positive tubercle bacilli that had been phagocytosed by neutrophils. Ziehl–Neelsen staining, performed after the Gram staining, showed M. tuberculosis in the CSF (Fig. 2). The majority of M. tuberculosis organisms in the CSF were found inside neutrophils on Gram staining. The number of tubercle bacilli-phagocytic neutrophils was small (tubercle bacilli-phagocytic neutrophils/neutrophils/ 100 WBC in the CSF; 2/92/100) on Gram stain, and was one-half of that seen on Ziehl–Neelsen staining (data not shown). The small percentage of tubercle bacilli-phagocytic neutrophils indicates that careful attention must be paid to find bacilli in Gram-stained specimens. Our case showed for the first time that Gram staining is a useful tool to predict the presence of tuberculous meningitis. However, one possible limitation is that its usefulness may be restricted to cases in which the CSF shows predominantly neutrocytic pleocytosis but no other evidence of bacterial meningitis. Further studies are needed to clarify the usefulness of Gram staining in which the CSF shows lymphocytic pleocytosis. The ease of the procedure of the focus-changing technique when viewing Gram-stained specimens will contribute to its use as an initial diagnostic tool before confirming the diagnosis of tubercular meningitis with predominantly neutrocytic pleocytosis. This procedure would be especially useful in developing countries.
References 1. Thwaites GE, Chau TT, Mai NT, et al. Tuberculous meningitis. J Neurol Neurosurg Psychiatry. 2000;68:289–99. 2. Verdon R, Chevret S, Laissy JP, Wolff M. Tuberculous meningitis in adults: review of 48 cases. Clin Infect Dis. 1996;22: 982–8. 3. Atsukawa Y, Kawakami S, Asahara M, Ishigaki S, TanakaT, Ono Y, et al. The usefulness of focus changing examination in Gram staining as initial diagnostic clue for infective tuberculosis. J Infect Chemother. 2011;17:571–74. 4. Leonared JM, Des Prez RM. Tuberculous meningitis. Infect Dis Clin North Am. 1990;4:769–87. 5. Raviglione MC, O’Brien RJ. Tuberculosis (chapter 158) In: Harrison’s principles of internal medicine. 17th ed. New York: The McGraw-Hill Companies, Inc.; 2008. p. 2634. 6. Behlau I, Ellner JJ. Chronic meningitis (chapter 82) In: Principles and practice of infectious diseases. 6th ed. Philadelphia: Elsevier Churchill Livingstone; 2005. p. 1136. 7. Fitzgerald D, Haas D. Mycobacterium tuberculosis (chapter 248) In: Principles and practice of infectious diseases. 6th ed. Philadelphia: Elsevier Churchill Livingstone; 2005. p. 2877. 8. Raviglione MC, O’Brien RJ. Tuberculosis (chapter 158) In: Harrison’s principles of internal medicine. 17th ed. New York: The McGraw-Hill Companies, Inc.; 2008. p. 1012. 9. Choi SH, Kim YS, Bae IG, Chung JW. The possible role of cerebrospinal fluid adenosine deaminase activity in the diagnosis
123
936 of tuberculous meningitis in adults. Clin Neurol Neurosurg. 2002;104:10–5. 10. Hinson JM, Bradsher RW, Bodner SJ. Gram-stain neutrality of Mycobacterium tuberculosis. Am Rev Respir Dis. 1981;123: 365–6. 11. Trifiro S, Bourgault AM, Lebel F. Ghost mycobacteria on Gram stain. J Clin Microbiol. 1990;28:146–7.
123
J Infect Chemother (2012) 18:931–936 12. Atsukawa Y, Kawakami S, Ono Y, Fujisaki R, Miyazawa Y, Nishiya H. Three cases of Mycobacterium tuberculosis infection initially recognized by focus changing examination in gram staining. In: Roy PK, editor. Insight and control of infectious disease in global scenario. InTech—Open Access Company (ISBN 978-953-51-0319-6) (in press).