Background
Lyme disease caused by a spirochete
B. burgdorferi is the most commonly reported vector-borne disease in North America, and is transmitted by the blacklegged ticks,
I. scapularis and
Ixodes pacificus [
1]. In Canada, LD is emerging due to the spread of
I. scapularis and by 2020
, 80% of the population in south central and southeastern Canada could live in regions where LD risk areas occur [
2]. In 2004, there were an estimated 40 reported human cases of LD in Canada [
1]; this number rose to 917 in 2015 [
3]. The same trend can be observed in the province of Quebec with the number of reported human cases increasing from 2 in 2004 to 160 in 2015 [
4]. Most of the indigenous cases occurred after exposure to the Montérégie and Estrie regions [
4,
5].
Lyme disease is a multisystem infection that is manifested by progressive stages. In the early stage of the disease, which lasts from several days to several weeks, a cutaneous lesion, EM appears in approximately 70% of infected patients [
6]. The EM, which represents an inflammatory response to the spirochetes as they migrate through the skin away from the tick bite site, spreads to exceed 5 cm in most cases [
7]. Erythema migrans can be accompanied by flu-like symptoms such as fever, fatigue, headache, myalgia, or arthralgia [
6]. In many untreated cases,
B. burgdorferi disseminates throughout the body from the skin via the blood and causes manifestations of early disseminated LD [
8], which include neurological (e.g. facial palsy and meningitis), cardiac (heart block, which may on rare occasions be fatal), musculoskeletal and dermatologic (multiple secondary EM lesions) signs and symptoms [
8]. Again if untreated LD can progress to the late disseminated stage, during which arthritis is the most common manifestation [
9,
10].
Currently, serological testing using the two tier algorithm is the main laboratory diagnostic technique for LD in clinical practice [
11]. The two-step approach, comprising a first screening enzymatic immunoassay (EIA), followed by a confirmatory Western blot test [
7,
12], was developed to optimise specificity and sensitivity. However in early stages of LD, specific antibodies have not developed and serologic tests have poor sensitivity [
13]. Moreover, the sensitivity of the serological tests can be negatively affected in circumstances when the patient has been successfully treated with antibiotics. Hence, in the presence of EM and when the patient has a history of possible exposure to tick vectors, laboratory testing for LD antibodies is not recommended to support diagnosis [
7,
10]. The use of serologic testing is advocated to support diagnosis of early and late disseminated infection when the clinical signs can be confused with other diseases [
7].
Lyme disease diagnosis is a challenge for GPs in emerging areas. Previous studies that have described the practices and knowledge of physicians have shown that: 1) the majority of GPs lack knowledge to recognize the EM [
14]; 2) they do not know that EM alone is adequate for clinical diagnosis and 3) they tend to confirm the infection with laboratory tests [
15,
16], even though initiation of treatment without laboratory confirmation is usually appropriate [
17,
18].
In this 8-year retrospective study, data on the management of tick-bitten patients by GPs in Quebec were used to analyze their knowledge and practices regarding: 1) the recognition of early LD clinical signs; 2) appropriateness of prophylactic treatment; 3) treatment of LD cases and 4) laboratory tests requested. Evidence-based guidelines for the management of patients with LD, published in 2006 by the Infectious Diseases Society of America (IDSA) were used as the reference for recommended approaches to LD diagnosis and treatment [
10].
Discussion
Of those patients bitten by detectably
B. burgdorferi-infected
I. scapularis who were then diagnosed as having LD, almost one half were diagnosed as having EM less than 5 days after the tick was detected (and most of them were diagnosed at the moment of tick-removal) and we considered it likely that the skin lesions seen by the doctor were in fact reactions to the tick bite [
10,
21]. It was assumed that reported EM cases fulfilled other criteria for diagnosis of EM such as having a diameter of > 5 cm [
10]. A recent cross-sectional study has shown a moderate lack of knowledge of Lyme disease by family physicians in southern Quebec [
16], suggesting that increased physician awareness may reduce the likelihood of such misdiagnoses in the future.
In our study, more than half of the GPs are practicing in the Montérégie and Estrie health regions, where public health authorities have provided GPs with information on LD since 2010. The over-diagnosis evident in this study is probably due to the combination of lack of knowledge and high risk perception among these GPs. The possible negative consequences of over-diagnosis of EM are stress and adverse effects of antibiotics for the patient, an overuse of resources [
22], generating costs to the health system [
23], and probably a contribution to antibiotic resistance [
24,
25]. Evidence of lack of knowledge on Lyme disease is present despite efforts to provide information to practitioners suggests that some more novel methods of practitioner education may be needed.
However, in the other health regions of Quebec, where GPs have not yet received information to raise their awareness of LD, and where the risk for the population is lower than in Montérégie and Estrie health regions, EM may go unnoticed [
26]. An accurate diagnosis of early infection is sufficient to administer an antimicrobial treatment at a stage when the spread of the pathogen in the blood to other organs can, in most cases, be avoided [
10]. Not recognizing early LD clinical signs or delay in diagnosis could complicate the diagnosis and consequently require more intensive and longer-term antibiotic courses for the patient [
27].
The present study showed that few patients were diagnosed, by Quebec GPs, with EM beyond a month after exposure, but these could have been multiple EMs, which may appear up to 70 days post infection as part of the spectrum of manifestations of early disseminated LD [
21]. This result may indicate that Quebec GPs do recognize multiple EM as a manifestation of LD but further studies are needed to confirm this.
Previous estimates of risk of occurrence of LD after an infected
I. scapularis tick-bite range from 3.7% to 6% [
28,
29]. However, this study has shown that approximately one quarter of patients bitten by infected
I. scapularis were diagnosed as having LD cases by Quebec GPs. This difference could be explained by the over diagnosis of EM. Moreover, reports of confirmatory laboratory testing for the reported cases of disseminated LD in our study were scant and it is also possible that the number of disseminated LD cases was over-estimated due to neurological, cardiac or musculoskeletal manifestations caused by clinical conditions other than LD. When non-genuine EM and non-laboratory confirmed cases of disseminated LD were removed from the analysis, the proportion of bitten patients with infected
I. scapularis who developed LD was similar to that estimated by Huegli et al. [
29].
The results of this study showed that Quebec GPs are currently challenged in their practices regarding patients bitten by a tick in this area of LD emergence. The majority of prophylactic treatments of LD were not justified suggesting a lack of knowledge of recommended best practices [
10]. The unjustified prescription of antibiotic prophylaxis by Quebec GPs in the present study could be influenced by the intention to prevent LD or other infections transmitted by a tick bite or to yield to patients’ requests [
30,
31]. Furthermore, some prophylactic treatments were not aligned with the guidelines regarding the type of antibiotic prescribed and when the appropriate antibiotic was prescribed some therapeutic protocol did not follow the recommendations regarding the duration of treatment [
9,
10].
In contrast, this study showed that all the treatments prescribed for cases suffering LD were appropriate, regarding the choice of antimicrobial, as reported in other studies [
17,
18].
The diagnostic accuracy of serologic assays is dependent on multiple factors including timing of specimen collection with respect to disease state. For LD cases, as diagnosed by Quebec GPs, more than half of serological tests were requested less than 18 days from the date of tick removal. Serological evaluation for antibodies to
B. burgdorferi following removal of an attached tick or soon after EM is noticed is not helpful diagnostically – results are often negative and therefore of limited clinical utility. [
7,
12]. This tendency for GPs to attempt to confirm the diagnosis of EM by a laboratory test has been reported among clinicians in British Columbia and in the state of New Hampshire, USA [
15,
17]. The LD serologic testing is generally recommended in late LD when disseminated clinical signs could be confused with other diseases [
12].
For approximately one quarter of patients bitten by infected
I. scapularis who never showed clinical signs of LD during 3 months post exposure, one or more serological tests were requested. Serological tests should be used to help diagnosis in cases of disseminated LD, but in the absence of objective clinical signs, serological testing requests would be unjustified and are discouraged [
9,
10]. These findings suggest that Quebec GPs lack knowledge on appropriate practices regarding serologic testing for asymptomatic patient bitten by infected
I. scapularis as reported in Ferrouillet et al. [
16].
There are, of course, limitations to this study, the foremost of which is the representativeness of the sample. Data were collected by a simple questionnaire completed by GPs who received a positive
B. burgdorferi result for the analysis of a tick collected from the patient, and it could be argued that this test result influenced the practices of the GP regarding serological testing requests and treatment. In a recent study 69% of GPs in the study region considered that sending a tick taken from patient to the laboratory for identification and detection of
B. burgdorferi was useful for the diagnosis of LD [
16]. However, the tick identification and testing described here (and previously) is a surveillance tool and not a diagnostic procedure so there is 6–12 week delay between tick collection and return of test results. So the tick PCR test result would not have impacted practices of testing and treatments (including tick bite prophylactic treatments) prior to this. We also do not know if the practices of the GPs who completed the questionnaire are representative of the practices of all GPs in the province, and it could be argued that the GPs who participate in the surveillance, sent ticks for analysis and completed the questionnaire may be those with most awareness of Lyme disease. The time between tick removal and follow-up questionnaire varied from patient to patient which may have influenced the information available at the time of completion of the questionnaire. Moreover, the ≥5 days threshold for classifying an EM as a genuine EM was based on the date of tick removal without taking into account the time of tick attachment, however only a small proportion of ticks (16%) showed the signs of engorgement that would suggest they had fed for several days so we do not think we are overestimating the number of likely false EMs.
Conclusions
The present study showed that Quebec’s GPs are not familiar with early LD clinical signs and have tendency to confirm the clinical diagnosis by serologic testing when this is not recommended. Moreover, some GPs request serological tests for bitten patients even if no clinical sign was manifested. The majority of prophylactic treatments were not justified mainly because the removed tick hadn’t been attached to the patient long enough to transmit the causative agent of LD and also because most of the ticks were acquired in areas where the prevalence of B. burgdorferi was lower than 20%. However, the study showed appropriate practices regarding the type of antibiotic used for treatment of LD cases and diagnosis of multiple cutaneous manifestations.
Lyme disease is a public health concern that requires a multidisciplinary approach to reduce an increase in human incidence and to limit its impact in infected patients. It should be noted that for GPs to adhere to current guidelines requires that knowledge of the ecology of transmission cycles (specifically infection prevalence in ticks) is communicated to GPs and that the latter are trained to identify the species of the tick and its engorgement state.
Our study suggests that public health authorities need to target front-line health professionals with education on LD and on evidence-based guidelines for the management of exposed patients to avoid overuse of public health resources and for the benefit of the health of the population.
Acknowledgements
We thank the many people involved in passive tick surveillance over the past decades, and in particular the participating medical clinics and laboratories in Quebec. This work would not have been possible without the LSPQ and NML technical staff, Annie Rouleau, Alexandra Cecan, Marie Lemieux, Antonia Dibernardo, Tyler Cote, Philip Snarr and Katarina Bernat for their roles in the identification and diagnostic testing of ticks.