- Email: [email protected]
Capnocytophaga canimorsus infections in The Netherlands: a nationwide survey
ORIGINAL ARTICLE
EPIDEMIOLOGY
Capnocytophaga canimorsus infections in The Netherlands: a nationwide survey A. P. van Dam1,* and A. Jansz2 1) Department of Medical Microbiology, Leiden University Medical Centre, Leiden and 2) Department of Medical Microbiology, PAMM foundation, Veldhoven, The Netherlands
Abstract A retrospective nationwide survey on the occurrence of Capnocytophaga canimorsus and Capnocytopaga cynodegmi infections in The Netherlands over 3 years showed 32 cases, of which 31 were caused by C. canimorsus and one by an unspecified oxidase-positive Capnocytophaga strain. Twenty-eight patients had been diagnosed by blood culture, one by culture from both blood and cerebrospinal fluid (CSF), one by culture from a conjunctival swab, and two patients by 16S rRNA gene amplification by PCR directly from a blood or CSF specimen. The incidence rate was 0.67 infections per million population. Bacteraemia was found in 94% of the cases. The age range of patients was 38–80 years; 72% of them were male. Among 26 patients from whom clinical data were available, splenectomy was not reported, but alcoholism was reported in five. Nine patients (35%) had been admitted to the intensive-care unit, and three patients (13%) died. The mortality rate was much lower than observed in previous studies. Keywords: Capnocytophaga canimorsus, Capnocytophaga, epidemiology, outcome, sepsis Original Submission: 12 June 2009; Revised Submission: 9 February 2010; Accepted: 9 Feburary 2010 Editor: G. Pappas Article published online: 17 February 2010 Clin Microbiol Infect 2011; 17: 312–315 10.1111/j.1469-0691.2010.03195.x Corresponding author: A. P. van Dam, Department of Medical Microbiology, Leiden University Medical Centre, P.O. Box 9600, 2300 RC Leiden, The Netherlands E-mail: [email protected] *Present address: Department of Medical Microbiology, Onze Lieve Vrouwe Gasthuis, P.O. Box 95500, 1090 HM Amsterdam, The Netherlands
Introduction Capnocytophaga canimorsus occurs as a commensal in dogs’ mouths, but can give rise to a life-threatening infection in humans. Sepsis has been reported as the most common manifestation of infection with C. canimorsus [1], but other infections, such as meningitis [2,3], endocarditis [4], mycotic aneurysm [5], brain abscess [6] and endophthalmitis [7], have also been reported. The incidence of C. canimorsus sepsis was estimated to be 0.5 per million population in Denmark during the period 1982–1995, with a case-fatality rate of 31% [1]. A recently published study from California suggested an increase in the frequency of the disease during recent years [8]. Splenectomy and alcoholism have been reported as
important underlying conditions in C. canimorsus sepsis [9]. Most cases have been related to contact with dogs, but infection after cat bites has also been described [10]. A related species, Capnocytophaga cynodegmi, seems to be much less virulent. Only five C. cynodegmi infections in humans have been published, including one patient with a wound infection [11], one with continuous ambulant peritoneal dialysis peritonitis [12], two with cellulitis [13,14], one of whom had also bacteraemia, and only one patient with severe sepsis [15]. Positive oxidase, katalase and arginine dihydrolase tests distinguish the two species from other Capnocytophaga species, which occur as commensals in the human mouth and only cause sepsis in highly immunocompromised patients [11]. As the few studies on the epidemiology of the disease were published in the last decade of the last century, a nationwide study to investigate the number of C. canimorsus and C. cynodegmi infections in The Netherlands was performed.
Materials and Methods A questionnaire was sent to all 54 laboratories for clinical microbiology in The Netherlands. In the questionnaire, the
ª2010 The Authors Journal Compilation ª2010 European Society of Clinical Microbiology and Infectious Diseases
CMI
van Dam and Jansz
number of cases diagnosed in each laboratory in the period 2003–2005 was asked for. If cases had been diagnosed, data on sex and age of the patient, origin of the strain, possible risk factors, underlying diseases and outcome of the patient were requested. In addition, the National Reference Laboratory of Bacteriology was contacted to look for strains that had been submitted for identification by clinical laboratories from The Netherlands and that had been identified as C. canimorsus or C. cynodegmi. Laboratories that had sent the strain to the National Reference Laboratory were contacted for additional clinical information.
Results and Discussion Fifty-three of 54 laboratories returned the questionnaire. Of these laboratories, 19 had diagnosed at least one case of C. canimorsus infection during the 3-year study period. All together, 32 cases had been diagnosed (Table 1). Thirty-one strains had been cultured from 30 patients. In two additional patients, the diagnosis was established by direct sequencing of PCR-amplified 16S rDNA from a blood sample and a cerebrospinal fluid (CSF) sample, respectively, a method previously described [16,17]. Thirty of the isolates had been identified as C. canimorsus, 27 of them by 16S rRNA gene sequencing. C. cynodegmi had never been identified during these 3 years. One strain had only been identified as an oxidase-positive Capnocytophaga strain, and will be regarded as C. canimorsus in the clinical analysis. Of the strains, 29 had been cultured from blood, one from CSF, and one from a
Capnocytophaga canimorsus infections in The Netherlands
conjunctival swab. From one patient, a strain had been cultured from both blood and CSF. The number of C. canimorsus infections in The Netherlands was calculated as 32 cases among a population of 16 million in 3 years. The yearly incidence was therefore 0.67 per million, and the incidence of sepsis was 0.63 per million. This is not significantly different from the reported incidence in Denmark (0.5 per million population). In the region comprising the north and east of The Netherlands, which are less populated, the incidence was significantly higher than expected, whereas in the more populated region comprising the west and centre, the incidence was lower (v2 = 15.9, p 0.0004) (Fig. 1). Possible explanations could be a higher incidence of dog bites or a greater delay in patients seeing doctors in more rural areas. The mean age of the patients was 60 years (range, 38– 80 years). Nine patients were female, and 23 were male. A male predominance of this infection has been shown earlier [9], but in this study, the number of female patients was not significantly below 50%. However, because of the small number of cases, this frequency may be misleading. Data on risk factors and clinical outcome were reported for 26 patients. Underlying diseases were reported in nine patients. None of the patients was known to have had a previous splenectomy. Five patients had alcoholism, and in three patients other underlying conditions were mentioned: one had chronic obstructive pulmonary disease and low-dose prednisone, and two had diabetes mellitus, one of whom had
Northern and eastern region Observed cases: 19 Expected cases: 10 Population density: 288/km2
TABLE 1. Characteristics of patients with Capnocytophaga canimorsus infection Total number of cases Age (years), mean (range) Sex (male/female) Bacteriology Positive culture Blood CSF Eye swab Positive PCR (blood) Positive PCR (CSF) With clinical data Died ICU admission Animal contact Reported dog bite Wound licked by dog Reported cat bite Reported spider bite Risk factors Splenectomy Alcoholism Other risk factors (see text)
32 60 (38–81) 23/9 (v2 = 3.21, NS) 31a 29 1a 1 1b 1b 26 3 (12%) 9 (35%)c 13 (50%) 3 (12%) 0 1 (4%) 0 5 (19%) 3 (12%)
CSF, cerebrospinal fluid; ICU, intensive-care unit; NS, not significant. a From one patient, a positive culture was obtained from both CSF and blood. b 16S rRNA gene amplification. c Including two of the patients who subsequently died.
313
Western and central region Observed cases: 8 Expected cases: 15 Population density: 904/km2
Southern region Observed cases: 8 Expected cases: 6 Population density: 441/km2
Total number of cases = 35
FIG. 1. Observed and expected number of Capnocytophaga canimorsus infections in different regions of The Netherlands.
ª2010 The Authors Journal Compilation ª2010 European Society of Clinical Microbiology and Infectious Diseases, CMI, 17, 312–315
314
Clinical Microbiology and Infection, Volume 17 Number 2, February 2011
had a previous cerebrovascular accident and the other of whom had hypertension. A dog bite had been reported in 13 patients (50%). Three additional patients had been licked on a wound by a dog, and one patient had been scratched on the thorax by a dog. A cat bite was not reported in any of the patients. Interestingly, one patient reported a spider bite. This patient, a previously healthy 50-year old male, was admitted because of confusion and suspicion of bacterial meningitis. His C-reactive protein on admission was 345 mg/L. A lumbar puncture revealed numerous leukocytes in after Gram staining, but no bacteria. The patient was treated with penicillin G (3 million units, 6 times daily), and improved over the next few days. Gram-negative rods grew from the anaerobic blood culture after 2 days and from the aerobic blood culture bottle after 4 days. The patient improved gradually on continued penicillin G treatment, which was stopped after 7 days. At that moment, the patient could be discharged from the hospital in good condition; C-reactive protein was 8 mg/L. The isolate from blood was subsequently characterized as C. canimorsus by molecular methods. The only risk factor that could be identified was that the patient had been bitten by a spider while on holiday in Spain a month before; otherwise, the course of his disease was not remarkably different from that of the other patients in the study. An additional patient, who was not included in the study, because he developed the disease before the study period, reported a bat bite. It is known that C. canimorsus also occurs in other animals, such as cattle and sheep [18]. Further studies should elucidate whether bats and even insects, like spiders, are carriers of this bacterium. The mortality rate among the patients was 3/26 (13%), which is lower than reported in previous studies. A review of all 103 published cases up to 1996 showed a mortality rate of 30% [9]. In Denmark, the mortality rate in a nationwide study performed between 1982 and 1995 was 12/39 (31%) [1], and in California, a mortality rate of 30% (n = 30) was reported between 1972 and 2004 [8]. The three patients who died were 40, 45 and 69 years old, and were not among the oldest patients. Moreover, the oldest patient of the study, an 80-year-old man (patient 21), visited the outpatient department for chills and fever, was orally treated with amoxycillin–clavulanic acid, and had already recovered when the culture results became known. In our study, intensivecare unit admission was reported in nine patients (36%), including two of the three patients who subsequently died. Serious sequelae developed in two patients; in one of them, both legs had to be amputated because of disseminated intravascular coagulation. The low mortality rate in the present study may also be the result of the low number of underlying
CMI
diseases among the patients; especially after splenectomy, a severe course of the infection has been described [9]. One patient had a severe eye infection with C. canimorsus, possibly a superinfection after herpes keratitis. In January 2005, this patient had already been treated for presumed herpes keratitis with valacyclovir and local therapy with polyspectran and corticosteroids. In April 2005, the patient still had blurred vision, and C. canimorsus was cultured from a corneal swab of the affected eye. Prolonged therapy with chloramphenicol and ciprofloxacin locally and clindamycin orally was not successful. After several months, the affected eye was enucleated because of prolonged inflammation and blurred vision. The patient had two cats and no dogs; however, no direct contact between the cats and the patient’s eye had been reported. A similar case has recently been described in The Netherlands [7]. It should be acknowledged that a retrospective survey has a number of potential pitfalls. Under-reporting by laboratories may have occurred. The laboratories that responded provide hospital care for approximately 98% of The Netherlands. Of course, cases may have been missed because of inadequate sampling of materials, incorrect determination of cultured strains, and inadequate reports by regional laboratories. To include all cases, the National Reference Laboratory, to which strains with identification problems can be sent, had also been contacted. This laboratory reported ten cases, of which seven had also been reported by the regional laboratories. Two laboratories had not yet responded, and reported these cases afterwards. Only one strain had not been found in the database of the laboratory where it was originally cultured. Therefore, under-reporting by regional laboratories does not seem to be a major problem. Almost all strains had been characterized by 16S rRNA gene sequencing, which might have some inadequacies in discriminating between C. canimorsus and C. cynodegmi [19], but is nonetheless one of the best available methods. Twenty-eight of 32 patients were diagnosed outside the hospitals where the study was performed; therefore, no direct access to patient charts was possible, and data on risk factors, underlying diseases and outcome had to be based on data obtained by the survey. In our questionnaire, we asked explicitly about splenectomy and alcoholism, conditions that have previously been described as associated with an increased risk for C. canimorsus septicaemia. For about half of the 17 patients in whom no underlying diseases were reported, it was explicitly stated in the questionnaires that the patient had no splenectomy or alcoholism. However, regarding the other patients, clinical microbiologists usually contact clinicians about these uncommon pathogens, and would probably have noted important risk factors such as
ª2010 The Authors Journal Compilation ª2010 European Society of Clinical Microbiology and Infectious Diseases, CMI, 17, 312–315
CMI
van Dam and Jansz
those mentioned above. Unfortunately, we did not ask for data on antibiotic usage, on the initial diagnosis of the patient, and on when the correct diagnosis had been made. These factors might all have influenced the outcome of the disease. It is interesting that the number of patients with clearly identifiable underlying diseases was low. The sex and age distribution of the infected patients corresponded to those in earlier reports. Whereas a male predominance among the patients could be related to an increased risk for dog bites in men, the exclusive occurrence of the disease in middleaged and elderly persons points at waning immunity as a risk factor for the disease. Possible virulence factors for the species have recently been described, such as resistance to complement and polymorphonuclear leukocytes [20], and blocking of release of nitric oxide by lipopolysaccharide-activated macrophages [21], an ability not shared by most strains isolated from dogs [22]. The low mortality rate in the present study could well be a consequence of improvements in hospital and ICU facilities. Further and more extensive studies should confirm these hypotheses, and might clarify more immune mechanisms behind this rare, but often serious, infectious disease.
Acknowledgements We would like to thank all medical microbiologists who contributed to this study, and especially B. Overbeek, J. Weel, E. Roelofsen, S. Debast, A. Van Griethuysen, R. Schouten, K. Brinkman, H. Bijlmer, I. Frenay, H. Dessel, C. Verduin, F. Stals, F. Sebens, P. Wever, F. Heilmann, F. Reubsaet and R. Klont, who provided positive microbiological, clinical and epidemiological data on the patients in the present study.
Transparency Declaration The authors have no relationship (commercial or otherwise) that may constitute a dual or conflicting interest.
References 1. Pers C, Gahrn-Hansen B, Frederiksen W. Capnocytophaga canimorsus septicemia in Denmark, 1982–1995: review of 39 cases. Clin Infect Dis 1996; 23: 71–75. 2. de Boer MG, Lambregts PC, van Dam AP, van‘t Wout JW. Meningitis caused by Capnocytophaga canimorsus: when to expect the unexpected. Clin Neurol Neurosurg 2007; 109: 393–398.
Capnocytophaga canimorsus infections in The Netherlands
315
3. Gasch O, Fernandez N, Armisen A, Verdaguer R, Fernandez P. Community-acquired Capnocytophaga canimorsus meningitis in adults: report of one case with a subacute course and deafness, and literature review. Enferm Infecc Microbiol Clin 2009; 27: 33–36. 4. Ngaage DL, Kotidis KN, Sandoe JA, Unnikrishnan NR. Do not snog the dog: infective endocarditis due to Capnocytophaga canimorsus. Eur J Cardiothorac Surg 1999; 16: 362–363. 5. Tierney DM, Strauss LP, Sanchez JL. Capnocytophaga canimorsus mycotic abdominal aortic aneurysm: why the mailman is afraid of dogs. J Clin Microbiol 2006; 44: 649–651. 6. Ulivieri S, Oliveri G, Filosomi G. A case of Capnocytophaga canimorsus brain abscess secondary to dog’s bite. G Chir 2008; 29: 79–80. 7. Papadaki TG, el Moussaoui R, van Ketel RJ, Verbraak FD, Tan HS. Capnocytophaga canimorsus endogenous endophthalmitis in an immunocompetent host. Br J Ophthalmol 2008; 92: 1566–1567. 8. Janda JM, Graves MH, Lindquist D, Probert WS. Diagnosing Capnocytophaga canimorsus infections. Emerg Infect Dis 2006; 12: 340–342. 9. Lion C, Escande F, Burdin JC. Capnocytophaga canimorsus infections in human: review of the literature and cases report. Eur J Epidemiol 1996; 12: 521–533. 10. Valtonen M, Lauhio A, Carlson P et al. Capnocytophaga canimorsus septicemia: fifth report of a cat-associated infection and five other cases. Eur J Clin Microbiol Infect Dis 1995; 14: 520–523. 11. Brenner DJ, Hollis DG, Fanning GR, Weaver RE. Capnocytophaga canimorsus sp. nov. (formerly CDC group DF-2), a cause of septicemia following dog bite, and C. cynodegmi sp. nov., a cause of localized wound infection following dog bite. J Clin Microbiol 1989; 27: 231–235. 12. Pers C, Tvedegaard E, Christensen JJ, Bangsborg J. Capnocytophaga cynodegmi peritonitis in a peritoneal dialysis patient. J Clin Microbiol 2007; 45: 3844–3846. 13. Sarma PS, Mohanty S. Capnocytophaga cynodegmi cellulitis, bacteremia, and pneumonitis in a diabetic man. J Clin Microbiol 2001; 39: 2028– 2029. 14. Gerster JC, Dudler J. Cellulitis caused by Capnocytophaga cynodegmi associated with etanercept treatment in a patient with rheumatoid arthritis. Clin Rheumatol 2004; 23: 570–571. 15. Khawari AA, Myers JW, Ferguson DA Jr, Moorman JP. Sepsis and meningitis due to Capnocytophaga cynodegmi after splenectomy. Clin Infect Dis 2005; 40: 1709–1710. 16. Gottwein J, Zbinden R, Maibach RC, Herren T. Etiologic diagnosis of Capnocytophaga canimorsus meningitis by broad-range PCR. Eur J Clin Microbiol Infect Dis 2006; 25: 132–134. 17. Meybeck A, Aoun N, Granados D, Pease S, Yeni P. Meningitis due to Capnocytophaga canimorsus: contribution of 16S RNA ribosomal sequencing for species identification. Scand J Infect Dis 2006; 38: 375–377. 18. Westwell AJ, Kerr K, Spencer MB, Hutchinson DN. DF-2 infection. BMJ 1989; 298: 116–117. 19. van Dam AP, van Weert A, Harmanus C, Hovius KE, Claas EC, Reubsaet FA. Molecular characterization of Capnocytophaga canimorsus and other canine Capnocytophaga spp. and assessment by PCR of their frequencies in dogs. J Clin Microbiol 2009; 47: 3218–3225. 20. Shin H, Mally M, Meyer S et al. Resistance of Capnocytophaga canimorsus to killing by human complement and polymorphonuclear leukocytes. Infect Immun 2009; 77: 2262–2271. 21. Meyer S, Shin H, Cornelis GR. Capnocytophaga canimorsus resists phagocytosis by macrophages and blocks the ability of macrophages to kill other bacteria. Immunobiology 2008; 10: 805–814. 22. Mally M, Paroz C, Shin H et al. Prevalence of Capnocytophaga canimorsus in dogs and occurrence of potential virulence factors. Microbes Infect 2009; 11: 509–514.
ª2010 The Authors Journal Compilation ª2010 European Society of Clinical Microbiology and Infectious Diseases, CMI, 17, 312–315
EPIDEMIOLOGY
Capnocytophaga canimorsus infections in The Netherlands: a nationwide survey A. P. van Dam1,* and A. Jansz2 1) Department of Medical Microbiology, Leiden University Medical Centre, Leiden and 2) Department of Medical Microbiology, PAMM foundation, Veldhoven, The Netherlands
Abstract A retrospective nationwide survey on the occurrence of Capnocytophaga canimorsus and Capnocytopaga cynodegmi infections in The Netherlands over 3 years showed 32 cases, of which 31 were caused by C. canimorsus and one by an unspecified oxidase-positive Capnocytophaga strain. Twenty-eight patients had been diagnosed by blood culture, one by culture from both blood and cerebrospinal fluid (CSF), one by culture from a conjunctival swab, and two patients by 16S rRNA gene amplification by PCR directly from a blood or CSF specimen. The incidence rate was 0.67 infections per million population. Bacteraemia was found in 94% of the cases. The age range of patients was 38–80 years; 72% of them were male. Among 26 patients from whom clinical data were available, splenectomy was not reported, but alcoholism was reported in five. Nine patients (35%) had been admitted to the intensive-care unit, and three patients (13%) died. The mortality rate was much lower than observed in previous studies. Keywords: Capnocytophaga canimorsus, Capnocytophaga, epidemiology, outcome, sepsis Original Submission: 12 June 2009; Revised Submission: 9 February 2010; Accepted: 9 Feburary 2010 Editor: G. Pappas Article published online: 17 February 2010 Clin Microbiol Infect 2011; 17: 312–315 10.1111/j.1469-0691.2010.03195.x Corresponding author: A. P. van Dam, Department of Medical Microbiology, Leiden University Medical Centre, P.O. Box 9600, 2300 RC Leiden, The Netherlands E-mail: [email protected] *Present address: Department of Medical Microbiology, Onze Lieve Vrouwe Gasthuis, P.O. Box 95500, 1090 HM Amsterdam, The Netherlands
Introduction Capnocytophaga canimorsus occurs as a commensal in dogs’ mouths, but can give rise to a life-threatening infection in humans. Sepsis has been reported as the most common manifestation of infection with C. canimorsus [1], but other infections, such as meningitis [2,3], endocarditis [4], mycotic aneurysm [5], brain abscess [6] and endophthalmitis [7], have also been reported. The incidence of C. canimorsus sepsis was estimated to be 0.5 per million population in Denmark during the period 1982–1995, with a case-fatality rate of 31% [1]. A recently published study from California suggested an increase in the frequency of the disease during recent years [8]. Splenectomy and alcoholism have been reported as
important underlying conditions in C. canimorsus sepsis [9]. Most cases have been related to contact with dogs, but infection after cat bites has also been described [10]. A related species, Capnocytophaga cynodegmi, seems to be much less virulent. Only five C. cynodegmi infections in humans have been published, including one patient with a wound infection [11], one with continuous ambulant peritoneal dialysis peritonitis [12], two with cellulitis [13,14], one of whom had also bacteraemia, and only one patient with severe sepsis [15]. Positive oxidase, katalase and arginine dihydrolase tests distinguish the two species from other Capnocytophaga species, which occur as commensals in the human mouth and only cause sepsis in highly immunocompromised patients [11]. As the few studies on the epidemiology of the disease were published in the last decade of the last century, a nationwide study to investigate the number of C. canimorsus and C. cynodegmi infections in The Netherlands was performed.
Materials and Methods A questionnaire was sent to all 54 laboratories for clinical microbiology in The Netherlands. In the questionnaire, the
ª2010 The Authors Journal Compilation ª2010 European Society of Clinical Microbiology and Infectious Diseases
CMI
van Dam and Jansz
number of cases diagnosed in each laboratory in the period 2003–2005 was asked for. If cases had been diagnosed, data on sex and age of the patient, origin of the strain, possible risk factors, underlying diseases and outcome of the patient were requested. In addition, the National Reference Laboratory of Bacteriology was contacted to look for strains that had been submitted for identification by clinical laboratories from The Netherlands and that had been identified as C. canimorsus or C. cynodegmi. Laboratories that had sent the strain to the National Reference Laboratory were contacted for additional clinical information.
Results and Discussion Fifty-three of 54 laboratories returned the questionnaire. Of these laboratories, 19 had diagnosed at least one case of C. canimorsus infection during the 3-year study period. All together, 32 cases had been diagnosed (Table 1). Thirty-one strains had been cultured from 30 patients. In two additional patients, the diagnosis was established by direct sequencing of PCR-amplified 16S rDNA from a blood sample and a cerebrospinal fluid (CSF) sample, respectively, a method previously described [16,17]. Thirty of the isolates had been identified as C. canimorsus, 27 of them by 16S rRNA gene sequencing. C. cynodegmi had never been identified during these 3 years. One strain had only been identified as an oxidase-positive Capnocytophaga strain, and will be regarded as C. canimorsus in the clinical analysis. Of the strains, 29 had been cultured from blood, one from CSF, and one from a
Capnocytophaga canimorsus infections in The Netherlands
conjunctival swab. From one patient, a strain had been cultured from both blood and CSF. The number of C. canimorsus infections in The Netherlands was calculated as 32 cases among a population of 16 million in 3 years. The yearly incidence was therefore 0.67 per million, and the incidence of sepsis was 0.63 per million. This is not significantly different from the reported incidence in Denmark (0.5 per million population). In the region comprising the north and east of The Netherlands, which are less populated, the incidence was significantly higher than expected, whereas in the more populated region comprising the west and centre, the incidence was lower (v2 = 15.9, p 0.0004) (Fig. 1). Possible explanations could be a higher incidence of dog bites or a greater delay in patients seeing doctors in more rural areas. The mean age of the patients was 60 years (range, 38– 80 years). Nine patients were female, and 23 were male. A male predominance of this infection has been shown earlier [9], but in this study, the number of female patients was not significantly below 50%. However, because of the small number of cases, this frequency may be misleading. Data on risk factors and clinical outcome were reported for 26 patients. Underlying diseases were reported in nine patients. None of the patients was known to have had a previous splenectomy. Five patients had alcoholism, and in three patients other underlying conditions were mentioned: one had chronic obstructive pulmonary disease and low-dose prednisone, and two had diabetes mellitus, one of whom had
Northern and eastern region Observed cases: 19 Expected cases: 10 Population density: 288/km2
TABLE 1. Characteristics of patients with Capnocytophaga canimorsus infection Total number of cases Age (years), mean (range) Sex (male/female) Bacteriology Positive culture Blood CSF Eye swab Positive PCR (blood) Positive PCR (CSF) With clinical data Died ICU admission Animal contact Reported dog bite Wound licked by dog Reported cat bite Reported spider bite Risk factors Splenectomy Alcoholism Other risk factors (see text)
32 60 (38–81) 23/9 (v2 = 3.21, NS) 31a 29 1a 1 1b 1b 26 3 (12%) 9 (35%)c 13 (50%) 3 (12%) 0 1 (4%) 0 5 (19%) 3 (12%)
CSF, cerebrospinal fluid; ICU, intensive-care unit; NS, not significant. a From one patient, a positive culture was obtained from both CSF and blood. b 16S rRNA gene amplification. c Including two of the patients who subsequently died.
313
Western and central region Observed cases: 8 Expected cases: 15 Population density: 904/km2
Southern region Observed cases: 8 Expected cases: 6 Population density: 441/km2
Total number of cases = 35
FIG. 1. Observed and expected number of Capnocytophaga canimorsus infections in different regions of The Netherlands.
ª2010 The Authors Journal Compilation ª2010 European Society of Clinical Microbiology and Infectious Diseases, CMI, 17, 312–315
314
Clinical Microbiology and Infection, Volume 17 Number 2, February 2011
had a previous cerebrovascular accident and the other of whom had hypertension. A dog bite had been reported in 13 patients (50%). Three additional patients had been licked on a wound by a dog, and one patient had been scratched on the thorax by a dog. A cat bite was not reported in any of the patients. Interestingly, one patient reported a spider bite. This patient, a previously healthy 50-year old male, was admitted because of confusion and suspicion of bacterial meningitis. His C-reactive protein on admission was 345 mg/L. A lumbar puncture revealed numerous leukocytes in after Gram staining, but no bacteria. The patient was treated with penicillin G (3 million units, 6 times daily), and improved over the next few days. Gram-negative rods grew from the anaerobic blood culture after 2 days and from the aerobic blood culture bottle after 4 days. The patient improved gradually on continued penicillin G treatment, which was stopped after 7 days. At that moment, the patient could be discharged from the hospital in good condition; C-reactive protein was 8 mg/L. The isolate from blood was subsequently characterized as C. canimorsus by molecular methods. The only risk factor that could be identified was that the patient had been bitten by a spider while on holiday in Spain a month before; otherwise, the course of his disease was not remarkably different from that of the other patients in the study. An additional patient, who was not included in the study, because he developed the disease before the study period, reported a bat bite. It is known that C. canimorsus also occurs in other animals, such as cattle and sheep [18]. Further studies should elucidate whether bats and even insects, like spiders, are carriers of this bacterium. The mortality rate among the patients was 3/26 (13%), which is lower than reported in previous studies. A review of all 103 published cases up to 1996 showed a mortality rate of 30% [9]. In Denmark, the mortality rate in a nationwide study performed between 1982 and 1995 was 12/39 (31%) [1], and in California, a mortality rate of 30% (n = 30) was reported between 1972 and 2004 [8]. The three patients who died were 40, 45 and 69 years old, and were not among the oldest patients. Moreover, the oldest patient of the study, an 80-year-old man (patient 21), visited the outpatient department for chills and fever, was orally treated with amoxycillin–clavulanic acid, and had already recovered when the culture results became known. In our study, intensivecare unit admission was reported in nine patients (36%), including two of the three patients who subsequently died. Serious sequelae developed in two patients; in one of them, both legs had to be amputated because of disseminated intravascular coagulation. The low mortality rate in the present study may also be the result of the low number of underlying
CMI
diseases among the patients; especially after splenectomy, a severe course of the infection has been described [9]. One patient had a severe eye infection with C. canimorsus, possibly a superinfection after herpes keratitis. In January 2005, this patient had already been treated for presumed herpes keratitis with valacyclovir and local therapy with polyspectran and corticosteroids. In April 2005, the patient still had blurred vision, and C. canimorsus was cultured from a corneal swab of the affected eye. Prolonged therapy with chloramphenicol and ciprofloxacin locally and clindamycin orally was not successful. After several months, the affected eye was enucleated because of prolonged inflammation and blurred vision. The patient had two cats and no dogs; however, no direct contact between the cats and the patient’s eye had been reported. A similar case has recently been described in The Netherlands [7]. It should be acknowledged that a retrospective survey has a number of potential pitfalls. Under-reporting by laboratories may have occurred. The laboratories that responded provide hospital care for approximately 98% of The Netherlands. Of course, cases may have been missed because of inadequate sampling of materials, incorrect determination of cultured strains, and inadequate reports by regional laboratories. To include all cases, the National Reference Laboratory, to which strains with identification problems can be sent, had also been contacted. This laboratory reported ten cases, of which seven had also been reported by the regional laboratories. Two laboratories had not yet responded, and reported these cases afterwards. Only one strain had not been found in the database of the laboratory where it was originally cultured. Therefore, under-reporting by regional laboratories does not seem to be a major problem. Almost all strains had been characterized by 16S rRNA gene sequencing, which might have some inadequacies in discriminating between C. canimorsus and C. cynodegmi [19], but is nonetheless one of the best available methods. Twenty-eight of 32 patients were diagnosed outside the hospitals where the study was performed; therefore, no direct access to patient charts was possible, and data on risk factors, underlying diseases and outcome had to be based on data obtained by the survey. In our questionnaire, we asked explicitly about splenectomy and alcoholism, conditions that have previously been described as associated with an increased risk for C. canimorsus septicaemia. For about half of the 17 patients in whom no underlying diseases were reported, it was explicitly stated in the questionnaires that the patient had no splenectomy or alcoholism. However, regarding the other patients, clinical microbiologists usually contact clinicians about these uncommon pathogens, and would probably have noted important risk factors such as
ª2010 The Authors Journal Compilation ª2010 European Society of Clinical Microbiology and Infectious Diseases, CMI, 17, 312–315
CMI
van Dam and Jansz
those mentioned above. Unfortunately, we did not ask for data on antibiotic usage, on the initial diagnosis of the patient, and on when the correct diagnosis had been made. These factors might all have influenced the outcome of the disease. It is interesting that the number of patients with clearly identifiable underlying diseases was low. The sex and age distribution of the infected patients corresponded to those in earlier reports. Whereas a male predominance among the patients could be related to an increased risk for dog bites in men, the exclusive occurrence of the disease in middleaged and elderly persons points at waning immunity as a risk factor for the disease. Possible virulence factors for the species have recently been described, such as resistance to complement and polymorphonuclear leukocytes [20], and blocking of release of nitric oxide by lipopolysaccharide-activated macrophages [21], an ability not shared by most strains isolated from dogs [22]. The low mortality rate in the present study could well be a consequence of improvements in hospital and ICU facilities. Further and more extensive studies should confirm these hypotheses, and might clarify more immune mechanisms behind this rare, but often serious, infectious disease.
Acknowledgements We would like to thank all medical microbiologists who contributed to this study, and especially B. Overbeek, J. Weel, E. Roelofsen, S. Debast, A. Van Griethuysen, R. Schouten, K. Brinkman, H. Bijlmer, I. Frenay, H. Dessel, C. Verduin, F. Stals, F. Sebens, P. Wever, F. Heilmann, F. Reubsaet and R. Klont, who provided positive microbiological, clinical and epidemiological data on the patients in the present study.
Transparency Declaration The authors have no relationship (commercial or otherwise) that may constitute a dual or conflicting interest.
References 1. Pers C, Gahrn-Hansen B, Frederiksen W. Capnocytophaga canimorsus septicemia in Denmark, 1982–1995: review of 39 cases. Clin Infect Dis 1996; 23: 71–75. 2. de Boer MG, Lambregts PC, van Dam AP, van‘t Wout JW. Meningitis caused by Capnocytophaga canimorsus: when to expect the unexpected. Clin Neurol Neurosurg 2007; 109: 393–398.
Capnocytophaga canimorsus infections in The Netherlands
315
3. Gasch O, Fernandez N, Armisen A, Verdaguer R, Fernandez P. Community-acquired Capnocytophaga canimorsus meningitis in adults: report of one case with a subacute course and deafness, and literature review. Enferm Infecc Microbiol Clin 2009; 27: 33–36. 4. Ngaage DL, Kotidis KN, Sandoe JA, Unnikrishnan NR. Do not snog the dog: infective endocarditis due to Capnocytophaga canimorsus. Eur J Cardiothorac Surg 1999; 16: 362–363. 5. Tierney DM, Strauss LP, Sanchez JL. Capnocytophaga canimorsus mycotic abdominal aortic aneurysm: why the mailman is afraid of dogs. J Clin Microbiol 2006; 44: 649–651. 6. Ulivieri S, Oliveri G, Filosomi G. A case of Capnocytophaga canimorsus brain abscess secondary to dog’s bite. G Chir 2008; 29: 79–80. 7. Papadaki TG, el Moussaoui R, van Ketel RJ, Verbraak FD, Tan HS. Capnocytophaga canimorsus endogenous endophthalmitis in an immunocompetent host. Br J Ophthalmol 2008; 92: 1566–1567. 8. Janda JM, Graves MH, Lindquist D, Probert WS. Diagnosing Capnocytophaga canimorsus infections. Emerg Infect Dis 2006; 12: 340–342. 9. Lion C, Escande F, Burdin JC. Capnocytophaga canimorsus infections in human: review of the literature and cases report. Eur J Epidemiol 1996; 12: 521–533. 10. Valtonen M, Lauhio A, Carlson P et al. Capnocytophaga canimorsus septicemia: fifth report of a cat-associated infection and five other cases. Eur J Clin Microbiol Infect Dis 1995; 14: 520–523. 11. Brenner DJ, Hollis DG, Fanning GR, Weaver RE. Capnocytophaga canimorsus sp. nov. (formerly CDC group DF-2), a cause of septicemia following dog bite, and C. cynodegmi sp. nov., a cause of localized wound infection following dog bite. J Clin Microbiol 1989; 27: 231–235. 12. Pers C, Tvedegaard E, Christensen JJ, Bangsborg J. Capnocytophaga cynodegmi peritonitis in a peritoneal dialysis patient. J Clin Microbiol 2007; 45: 3844–3846. 13. Sarma PS, Mohanty S. Capnocytophaga cynodegmi cellulitis, bacteremia, and pneumonitis in a diabetic man. J Clin Microbiol 2001; 39: 2028– 2029. 14. Gerster JC, Dudler J. Cellulitis caused by Capnocytophaga cynodegmi associated with etanercept treatment in a patient with rheumatoid arthritis. Clin Rheumatol 2004; 23: 570–571. 15. Khawari AA, Myers JW, Ferguson DA Jr, Moorman JP. Sepsis and meningitis due to Capnocytophaga cynodegmi after splenectomy. Clin Infect Dis 2005; 40: 1709–1710. 16. Gottwein J, Zbinden R, Maibach RC, Herren T. Etiologic diagnosis of Capnocytophaga canimorsus meningitis by broad-range PCR. Eur J Clin Microbiol Infect Dis 2006; 25: 132–134. 17. Meybeck A, Aoun N, Granados D, Pease S, Yeni P. Meningitis due to Capnocytophaga canimorsus: contribution of 16S RNA ribosomal sequencing for species identification. Scand J Infect Dis 2006; 38: 375–377. 18. Westwell AJ, Kerr K, Spencer MB, Hutchinson DN. DF-2 infection. BMJ 1989; 298: 116–117. 19. van Dam AP, van Weert A, Harmanus C, Hovius KE, Claas EC, Reubsaet FA. Molecular characterization of Capnocytophaga canimorsus and other canine Capnocytophaga spp. and assessment by PCR of their frequencies in dogs. J Clin Microbiol 2009; 47: 3218–3225. 20. Shin H, Mally M, Meyer S et al. Resistance of Capnocytophaga canimorsus to killing by human complement and polymorphonuclear leukocytes. Infect Immun 2009; 77: 2262–2271. 21. Meyer S, Shin H, Cornelis GR. Capnocytophaga canimorsus resists phagocytosis by macrophages and blocks the ability of macrophages to kill other bacteria. Immunobiology 2008; 10: 805–814. 22. Mally M, Paroz C, Shin H et al. Prevalence of Capnocytophaga canimorsus in dogs and occurrence of potential virulence factors. Microbes Infect 2009; 11: 509–514.
ª2010 The Authors Journal Compilation ª2010 European Society of Clinical Microbiology and Infectious Diseases, CMI, 17, 312–315