Incidence and antibiotic treatment of erythema ... - Legeforeningen

G Model TTBDIS-684; No. of Pages 8

ARTICLE IN PRESS Ticks and Tick-borne Diseases xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Ticks and Tick-borne Diseases journal homepage: www.elsevier.com/locate/ttbdis

Incidence and antibiotic treatment of erythema migrans in Norway 2005–2009 Knut Eirik Eliassen a,∗ , Dag Berild b , Harald Reiso c , Nils Grude d , Karen Sofie Christophersen e , Cecilie Finckenhagen f , Morten Lindbæk a a

Norwegian Antibiotic Centre for Primary Care, Department of General Practice, University of Oslo, PO box 1130 Blindern, N-0318 Oslo, Norway Department of Infectious Diseases, Oslo University Hospital, Faculty of Medicine, University of Oslo, Kirkeveien 166, N-0460 Oslo, Norway c Norwegian National Advisory Unit on Tick-borne Diseases, Sorlandet Hospital, PO box 783 Stoa, N-4809 Arendal, Norway d Department of Medical Microbiology, Vestfold Hospital Trust, Halfdan Wilhelmsens allè 17, N-3116 Tonsberg, Norway e Tarnasen GP Office, Valhallaveien 70, N-1413 Tarnasen, Norway f Asker and Baerum Primary Care Out-of-hours Service, Sogneprest Munthe-kaas vei 100, N-1346 Gjettum, Norway b

a r t i c l e

i n f o

Article history: Received 8 April 2016 Received in revised form 28 June 2016 Accepted 29 June 2016 Available online xxx Keywords: Lyme borreliosis Tick-borne diseases Antibiotic use Epidemiology Clinic General practice

a b s t r a c t The first stage of Lyme borreliosis (LB) is mainly the typical skin lesion, erythema migrans (EM), which is estimated to comprise 80–90% of all LB cases. However, the reporting of, and the actual incidence of LB varies throughout Europe. Studies from Sweden and Holland have found EM incidences varying from 53 to 464 EM/100,000 inhabitants/year. Under-reporting of LB is common and a coefficient of three to reach a realistic estimate is suggested. In Norway, it is mandatory to report only the second and third LB stages to the National Institute of Public Health. To find the Norwegian incidence of EM, we extracted data from the electronic medical records of regular general practitioners and out-of-hours services in the four counties with the highest rates of registered LB in the 5 years from 2005 to 2009. We found an EM incidence of 448 EM/100,000 inhabitants/year in these counties, which yields a national incidence of 148 EM/100,000 inhabitants/year. Our findings show that solitary EMs comprised almost 96% of the total LB incidence in Norway. Older females have the highest rates of EM. Phenoxymethylpenicillin is the most commonly used drug to treat EM in Norway, which complies with the national guidelines for antibiotic use. Antibody tests are performed in 15% of cases. Less than 1% of patients are referred to secondary care. The study also shows a high number of patients seeking care for tick bites without signs of infection and there is an overuse of antibiotics in these patients. © 2016 Elsevier GmbH. All rights reserved.

1. Introduction Lyme borreliosis (LB)1 is caused by the bacterium Borrelia burgdorferi sensu lato, which is transmitted through tick bites (TBs). LB can manifest with several different symptoms, which are traditionally divided into three disease stages. The first stage

∗ Corresponding author. E-mail addresses: [email protected] (K.E. Eliassen), [email protected] (D. Berild), [email protected] (H. Reiso), [email protected] (N. Grude), karen.sofi[email protected] (K.S. Christophersen), cecilie.fi[email protected] (C. Finckenhagen), [email protected] (M. Lindbæk) 1 Abbreviations: EM = erythema migrans, EMR = electronical medical record, LB = Lyme borreliosis, OOH = out-of-hours service, RGP = regular general practitioner, PCR = polymerase chain reaction, TB = tick bite

is mainly the skin lesion, erythema migrans (EM), which is estimated to comprise 80–90% of all LB cases. However, little is known about its actual incidence (Rizzoli et al., 2011; Stanek et al., 2012; Vandenesch et al., 2014; Hofhuis et al., 2015). In Norway, it is not mandatory to report solitary EMs. Only disseminated LB cases, confirmed by antibody testing, culture confirmation or polymerase chain reaction (PCR)2 analysis, are registered at the Norwegian Institute of Public Health. In Norway, the distribution of systemic manifestations in the study period (2005–2009) was 66.6% Lyme neuroborreliosis, 9.7% Lyme arthritis and 23.7% other or not specified manifestations, including Lyme carditis and multiple EMs (Personal communication, Myking S., Norwegian Institute of Public Health, 2016.06.02; MSIS, 2010).

2

PCR is a method to multiply and identify DNA from a tissue sample.

http://dx.doi.org/10.1016/j.ttbdis.2016.06.006 1877-959X/© 2016 Elsevier GmbH. All rights reserved.

Please cite this article in press as: Eliassen, K.E., et al., Incidence and antibiotic treatment of erythema migrans in Norway 2005–2009. Ticks Tick-borne Dis. (2016), http://dx.doi.org/10.1016/j.ttbdis.2016.06.006


ARTICLE IN PRESS K.E. Eliassen et al. / Ticks and Tick-borne Diseases xxx (2016) xxx–xxx

2

According to the European Concerted Action on Lyme Borreliosis (EUCALB), the incidence of LB in European countries is increasing. In 2006, the World Health Organization (WHO) reported that the incidence of LB varied from 0.6/100,000/year in Ireland to 155/100,000/year in Slovenia (O’Connell, 1996; Lindgren and Jaenson, 2006; EUCALB, 2015). In this report, LB incidence in Norway was reported as 2.8/100,000/year, but increased to a peak of 7.3 LB/100,000/year in 2008. In the United States (US), the incidence of LB in 2005–2009 was on average 8.9/100,000 inhabitants/year (CDC, 2015). The US figures include EM cases (CDC, 2011). Hubalek (2009) performed a thorough overview of LB epidemiology, and found 85,500 LB cases worldwide annually, with 65,500 of these cases in Europe. However, under-reporting is common, and multiplying the reported incidence by three is suggested to reach a realistic estimate of the number of LB cases (Hubalek, 2009). In Sweden, it is not mandatory to report LB, but it is estimated to be 69/100,000/year, of which EM comprises 77% of all LB cases (Berglund et al., 1995; Public Health Agency of Sweden, 2013). However, Bennet et al. (2006) found a yearly incidence of 464 EM/100,000 inhabitants in an endemic county in Sweden. Studies of EM incidence in European countries are scarce (Smith and Takkinen, 2006), but in Holland, Hofhuis et al. (2015) found an incidence of general practitioner (GP) consultations for TBs and EM of 495 and 132 per 100,000 inhabitants/year, respectively. The first aim of this study is to estimate the incidence of solitary EMs in Norway. The incidence of consultations for EMs in the four Norwegian counties having the highest reported incidence rates of systemic LB was registered to generate an estimate of the national incidence. The secondary aims of this study are to measure the compliance with national guidelines for antibiotic use, to assess the extent of referral to secondary care and antibody testing used. Doctor-seeking behaviour for mere TBs was also assessed. 2. Material and methods 2.1. Clinical setting We performed our study in the four Norwegian counties with the highest reported incidence rates of systemic LB in Norway in 2008. With a reported incidence of LB cases of 18.9/100,000 inhabitants/year, these counties (Vestfold, Telemark, Aust-Agder and Vest-Agder) had almost three times the national rate of 7.3 LB/100,000 inhabitants/year. The total population of the study counties was 669,739, comprising 14.3% of Norway’s total population in the study period (Statistics Norway, 2010). In Norway, all citizens are assigned a regular GP (RGP) as their primary contact in the health-care system. Outside office hours, each municipality is obliged to have a GP on call in an out-of-hours service (OOH) (Morken et al., 2009). Except for acute admissions, patients need a referral from a GP to be admitted to the hospital. Therefore, we assumed that nearly all patients with a TB or an EM in need of medical care will visit either their RGP or an OOH. 2.2. Inclusion of physicians We invited all RGPs and OOHs in the four study counties to participate in this study, which comprised 563 GPs and 25 OOHs. Limitations to the software used made it impossible to collect data from all different electronic medical record (EMR) systems (see Appendix). If the GP or OOH had an online server, data collection was not possible either. 407 GPs and 18 OOHs were able to participate and we obtained data from 213 (52%) and 14 (78%) of these participants, respectively (Fig. 1a). The number of patients on each GP’s list was retrieved from the Norwegian Health Economics Administration for the mid-date of each GP’s data collection period

(HELFO, 2016). For the OOHs, we used their catchment areas. The GPs had on average 1223 patients on their lists, and the 14 OOHs covered a population of 573,154 inhabitants.

2.3. Registration method In Norway, both RGPs and OOHs use the International Classification of Primary Care (ICPC-2) diagnostic coding system (Classification Committee of the World Organization of Family Doctors, 2011). “Erythema migrans” and “Tick bite” are specific diagnoses in the ICPC-2, but are included in broader ICPC codes like A78 Infectious disease and S73 Parasite infestation. Because of this, we believed that we would not identify all of the relevant cases by limiting the search to these diagnoses. Therefore, we developed a computer program that could search through the full text of the EMRs (see Appendix).

2.4. Data collection Data was collected in 2010 for the 5-year period from 1 January 2005 to 31 December 2009. The computer program identified 18 tick-related terms, including “erythema migrans” and “tick”, but also abbreviations and local Norwegian names for ticks. Spelling failures were included, as were Danish and Swedish words for ticks because many doctors from other Nordic countries work in Norway (see Appendix). When one or more of the search terms were matched, a dataset was extracted along with the full text of the journal note. The dataset included the following parameters: date of consultation, doctor’s ID number in the health personnel register, patient’s gender, age and postal code, diagnoses given and antibiotics prescribed. From the journal note, we found five additional parameters: whether the patient reported a recent or present TB, an EM, whether antibody testing had been performed, if the patient were referred to secondary care and whether the patient had been seen by another doctor prior to this consultation. We have only registered what the doctors documented, and have not interpreted clinical information to discern whether the diagnosis was correct or not.

2.5. Diagnostics and guidelines In the following, an “erythema migrans” is defined as a consultation for a solitary EM either with or without a known TB. It is the clinical diagnosis given by the GP based on the appearance of the skin lesion together with patient information about a TB or time spent in a tick-infested area (Stanek et al., 2011). We have only registered solitary EMs because multiple EMs are regarded as a systemic LB infection and should be treated and reported differently. A “tick bite” is defined as a consultation for a TB without an EM. Multiple consultations within 3 months of each other were counted as single cases. If a patient was first seen with a TB, and then came back with an EM, this was counted as an EM case. According to Norwegian guidelines, the antibiotic of choice for solitary EM is phenoxymethylpenicillin (PcV), with doxycycline and amoxicillin as alternatives. Antibiotic prophylaxis after TBs is not recommended in Norway (Norwegian Antibiotic Centre for Primary Care, 2015; Norwegian Directorate of Health, 2013; Nadelman et al., 2001). From the text search, we have registered whether antibody tests were performed or not. We have not been able to register the results of the tests.




3

Fig. 1. (a) Flowchart for inclusion of general practitioners (GPs) and out-of-hour services. (b) Number of patients in the different categories included in the study. EM is erythema migrans. “Rx” is antibiotic prescription. RGP is regular general practitioner. OOH is out-of-hour service.

2.6. Diagnoses given

2.8. Ethics

To evaluate the diagnoses given by the RGPs and OOHs, we scored them into four categories: relevant, related, caution and irrelevant. For details, see Appendix.

The patients were anonymised with patient numbers unique to the single dataset. Postal codes were used to identify the patients’ county of residence. The study was approved by the Regional Committee for Medical and Health Research Ethics (approval number 2009/1296; Regional Ethics Committee, 2010).

2.7. Statistics To estimate the national incidence of solitary EMs from our figures, we have assumed that the relation between solitary EMs and systemic LB, i.e., the “EM/LB factor”, in the study counties is the same as the relationship between solitary EMs and systemic LB nationally. The LB incidence was obtained from the Norwegian Surveillance System for Communicable Diseases (MSIS, 2010). EMincidence study counties EMincidence Norway = LBincidence study counties LBincidence Norway The data were analysed using the chi-square test for dichotomous variables and t-test for continuous variables, using p < 0.05 as significance level.

3. Results All figures are for RGP and OOH patients combined, unless otherwise stated. We retrieved data from 36,990 RGP and 13,555 OOH consultations. All consultation notes were read and scored based on the content of the text. This generated 8,134 RGP (22%) and 4,099 OOH (30%) consultations concerning a solitary EM or a mere TB. To compare our figures with national data, we have only included patients with a postal code within the four counties, which resulted in 7,667 (94%) and 3,492 (85%) of the registered RGP and OOH patients, respectively. Overlapping registrations between RGPs and OOHs were excluded. 253 (3.3%) of the RGP patients had already been seen at an OOH, and 85 (2.4%) of the OOH patients had already




4

been seen by an RGP. In addition, 111 (1.4%) of the RGP patients had already been consulted in secondary care prior to the registered consultation. For the following analyses, we used 6565 EM and 4256 TB consultations (Fig. 1b).

prescription vs. 0.2% (14/5,960) for the EMs with prescription. The rate of antibody testing was 54.2% (328/605) vs. 11.0% (654/5,960), respectively. The degree of uncertainty noted by the doctor was 23.9% (1,426/5,960) vs. 73.6% (445/605) for the EM patients who were prescribed antibiotics or not, respectively.

3.1. Age and gender distribution

3.5. Antibody testing

Of the EM patients, 3,455 (52.6%) were women. There was a significant male dominance in the 0–39-year-old age groups (51.9–59.0%; p < 0.001). In patients aged 40–80 years and over, there was a female dominance (51.8–58.8%). Most EMs were seen in the 50–69-year-old age groups (Fig. 2a). Among the patients with a TB only, there were 2145 (50.4%) men. There was a male dominance in the 0–49-year-old age groups (51.6–55.9%). In patients aged 50–80 years and over, there was a female dominance (52.7–55.5%). The youngest age group of 0–9 year olds had significantly more TB consultations than the other age groups, comprising 19.0% (810/4256) of all cases compared with an average of 10.1% in each age group (p < 0.05) (Fig. 2b).

Of all EM patients, 982 (15.0%) had an antibody test performed. At the RGPs, 935 (20.4%) had an antibody test while only 47 (2.4%) had an antibody test at the OOHs. Of the patients presenting with a TB without an EM, 606 (14.2%) had an antibody test performed. At the RGPs, 579 (20.5%) were tested and 27 (1.9%) of the patients at the OOHs were tested.

3.2. Incidence The mean incidence of consultations for solitary EMs in the four counties was 448 EM/100,000 inhabitants/year (Table 1). Considering the average national reported incidence of LB of 6.5/100,000 inhabitants/year and the average population of Norway (4,694,828) in the period of this study, we obtained an annual national estimate of EM incidence at 148 EM/100,000 inhabitants/year and an estimated national figure of 6,940 EM consultations/year. 79% of patients consulting their doctor for an EM did recall a recent TB prior to the EM. For consultations for mere TBs, we found an incidence of 284 TB consultations/100,000 inhabitants/year in the four counties. Assuming the same relationship between county and national incidence for TB consultations as for EM consultations, this gives us an estimated national incidence of (284*(148/448)) = 94 TB consultations/100,000 inhabitants/year or 4405 TB consultations/year.

3.6. Referral 56 (0.9%) of the EM patients and 43 (0.8%) of the TB only patients were referred to secondary care. The figures for the RGPs and OOHs were quite similar. 45 (1.0%) of the EM patients from the RGPs and 11 (0.6%) from the OOHs were referred. Rates for acute referral were 0.1% from the RGPs and 0.3% from the OOHs, respectively. 3.7. Diagnoses given For the four categories of patients, EM and TB at RGPs and OOHs, respectively, we have assessed which ICPC-2 diagnoses had been given (Table 2). Relevant diagnoses were given in 46.9% of the cases, and related diagnoses in 27.3% of the cases. Of the cases, 23.8% of the diagnoses given, were found to be irrelevant. TBs were given more relevant diagnoses than EMs, and the OOHs gave more relevant diagnoses than the RGPs. “Caution” diagnoses, indicating more severe tick-borne infection than an EM, were given in 2.0% of the cases. As patients could be given more than one diagnosis in the EMRs, it should be outlined that patients could be registered with two diagnoses, and one of them could be relevant; therefore, we obtained 12,415 diagnoses for 10,821 cases altogether. Where patients were given more than two diagnoses, the algorithm made sure that the relevant ones were registered (see Appendix).

3.3. Compliance with the guidelines 4. Discussion Altogether, 90.8% (5,090/6,565) of the EM cases were prescribed an antibiotic. Among patients presenting with a TB only, 842 (19.8%) received an antibiotic prescription. At the RGPs, 4031 (87.8%) of the EMs and 447 (15.8%) of the mere TBs were prescribed an antibiotic. For the OOHs, the rates were 1,929 (97.6%) and 395 (27.6%), respectively. Antibiotic prescribing for EM (Fig. 3a) was PcV in 3,932 (59.9%) of the cases, doxycycline in 1,733 (26.4%), macrolides in 137 (2.1%) and amoxicillin in 105 (1.6%). However, there was a higher prescribing rate for both macrolides 4.8% (31/649) and amoxicillin 5.9% (38/649) in the 0–9-year-old age group. Few other antibiotics were prescribed. In the “other” group, there were 28 prescriptions (0.4%) of antibiotics that did not fit into any of the other groups. The distribution of antibiotics was comparable in the EM and TB groups (Figs. 3b and c). 3.4. EM without prescription Of the EM patients, 9.2% (605/6565) were not prescribed an antibiotic. Therefore, we wanted to investigate whether these patients had received care elsewhere. At the time of registry, 11.7% (71/605) of those without a prescription had already been seen in secondary care. In comparison, only 0.4% (24/5960) of those who got a prescription had been seen elsewhere prior to the prescription. The referral rate was 6.9% (42/605) for the EMs without

4.1. Incidence This is the first study to show the incidence of consultations for solitary EMs and mere TBs in Norway. The main finding is the high and variable incidence of EM in the study counties leading to the national estimate of 148 EM consultations/100,000 inhabitants/year in Norway, and 94 consultations/100,000 inhabitants/year for mere TBs. It can be argued that these numbers are minimum figures because we have chosen only to use consultations at regular GPs with known list lengths. Some consultations may have occurred at these RGPs’ practices, but with their colleagues, an intern, a medical student or a locum tenens; therefore, these consultations are not registered by the method used here. Nevertheless, EM and TB accounted for at least 11,000 consultations in Norwegian primary care annually, which is much higher than the MSIS-registered 307 LB cases/year for the same period. This may partially explain the great attention that ticks and tick-borne diseases have been given in the media and among the general population (Stricker and Johnson, 2014). As the geographical distribution of ticks and Borrelia in Norway is very heterogeneous, one can argue that the incidence found in the four endemic counties is as relevant for clinicians in other endemic areas as is the national estimate.


G Model

ARTICLE IN PRESS

TTBDIS-684; No. of Pages 8

EM paents in study counes, n

a.

1400

60000

1200

50000

1000 40000 800 30000 600 20000 400

Populaon in study counes, n

K.E. Eliassen et al. / Ticks and Tick-borne Diseases xxx (2016) xxx–xxx

5

Female Male Male populaon Female populaon

10000

200 0

0

Tick bite paents in study counes, n

b.

1400

60000

1200

50000

1000 40000 800 30000 600 20000 400

Populaon in study counes, n

0-9 10-19 20-29 30-39 40-49 50-59 60-69 70-79 80+

Female Male Male populaon Female populaon

10000

200 0

0 0-9 10-19 20-29 30-39 40-49 50-59 60-69 70-79 80+

Fig. 2. (a) Age and gender distribution of erythema migrans and tick bite patients in Norwegian general practice 2005–2009. (a) Age and gender of erythema migrans patients, n = 6,565. (b) Age and gender of tick bite* ) patients, n = 4,256.* ) Patients with consultations for a tick bite without erythema migrans.

Table 1 Incidence of erythema migrans consultations (EM) in Norway 2005–2009. The estimated incidence of EM in Norway is calculated from the registered incidence of EM in four Norwegian counties and the incidence of systemic Lyme borreliosis (LB) nationally and in these counties. The incidence variation between the counties is given in the parentheses. EM incidence, LB incidence and population vary both within the counties and the years. Year

EM incidence counties n/100,000 inhabitants/year

LB incidence counties, MSIS1 n/100,000 inhabitants/year

EM/LB factor

LB incidence Norway, MSIS1 n/100,000 inhabitants/year

EM incidence Norway (estimated) n/100,000 inhabitants/year2

2005 2006 2007 2008 2009 Mean

353 (257–426) 552 (399–714) 450 (371–571) 455 (320–627) 431 (328–541) 448 (339–574)

22.1 (10.9–37.8) 22.6 (6.3–42.2) 21.4 (8.5–37.9) 18.9 (4.4–35.6) 13.4 (13.1–14.3) 19.7 (8.6–30.9)

16.0 24.4 21.0 24.1 32.1 22.7

6.1 6.7 7.0 7.3 5.7 6.5

97.6 164 147 176 183 97.6

1 Figures for the Lyme incidence in Norway in total and in the four counties are collected from the Norwegian Surveillance System for Communicable Diseases (MSIS, 2010). 2 Population data are from Statistics Norway (2010).

The figures are comparable to those in the Swedish and Dutch studies mentioned above (Bennet et al., 2006; Hofhuis et al., 2015). Counting about 7000 EM consultations per year in Norway, compared with an average of 307 reported LB cases/year in Norway, our findings show that solitary EMs comprised almost 96% of the total LB incidence in Norway.

4.2. Age and gender distribution Women aged over 60 years more frequently consult their doctor for TBs and EMs than do men, which has also been shown in previous studies (Bennet et al., 2007). The age and gender distribution for the TB consultations greatly resembles the pattern for EM, but the incidence peak in the youngest age group is not seen among the EM patients. However,


G Model

ARTICLE IN PRESS

TTBDIS-684; No. of Pages 8

K.E. Eliassen et al. / Ticks and Tick-borne Diseases xxx (2016) xxx–xxx

6

a.

1400 1200 1000 Unknown ab. 800

Other ab. Macrolides

600

Amoxicillin 400

Doxycycline

200

Phenoxymethylpenicillin

0 0-9

10-19 20-29 30-39 40-49 50-59 60-69 70-79

80+

Age group

b.

Phenoxymethylpenicillin Doxycycline Amoxicillin Macrolides Other ab. Unknown ab.

c.

Phenoxymethylpenicillin Doxycycline Amoxicillin Macrolides Other ab.

Fig. 3. (a) Antibiotic prescription for EM in Norwegian general practice 2005–2009, when prescribed, distribution in age groups, n = 5,960. (b) Distribution of antibiotics in EM patients, when prescribed, n = 5,960 (90,8% of 6,565). (c) Distribution of antibiotics in tick bite patients, when prescribed,n = 842 (19,8% of 4,256).

small children get Lyme neuroborreliosis more often than do older children and adults (Nygard et al., 2005). This may be because they get more TBs because of crawling and spending more time in the grass. They are also less capable of reporting TBs themselves and are more often bitten in the neck and head region (Tveitnes et al., 2012). From our figures, however, we cannot conclude that there are more actual TBs in these age groups, only that there is more doctor-seeking behaviour for TBs among the youngest children. The age and gender distribution of the whole population shows a small male preponderance of 51.1% in the age groups younger than 40 years, and a female dominance of 52% in the age groups older than 40 years (Figs. 2a and b). Neither the male dominance of EM-patients in the 40-year-old age groups of 55.6% (2541/4574) can be explained by the age and gender distribution in the population. In addition, the large number of consultations for TBs in the youngest age group cannot be explained from the age distribution.

4.3. Relationship between EMs and TBs A TB is not a disease, and most TBs do not lead to disease. In many cases, ticks bite people without their noticing. Most people who find an infested tick will probably remove it by themselves and not see a doctor unless they develop symptoms believed to be related to the TB. In this study, 79% of the EM patients did recall a recent TB, while in other studies of other LB manifestations, approximately half of the patients did recall a TB (Ljostad et al., 2003; Nadelman, 2015). The figure for TB consultations in this study of 94/100,000 inhabitants/year reflect the amount of people choosing to see their doctor after a TB without an EM. We expect the actual number of TBs to be much higher. 4.4. Antibiotic prescription In Norway, antibiotic prescribing complies with the antibiotic guidelines. About 60% receive PcV and more than 25% doxycycline.




7

Table 2 ICPC-2 diagnoses given for erythema migrans (EM) and tick bite patients in Norwegian general practice in 2005–2009. The “caution” category indicates systemic Lyme borreliosis or other tick borne disease. More than one diagnosis could be given in the electronic medical records, hence the n of 12,415 diagnoses for altogether 10,821 cases.

Macrolides are less effective against the Borrelia bacteria and are usually not recommended for LB treatment (Stanek et al., 2012), except for in children who are too young for doxycycline and in pregnant women in their second and third trimester who are allergic to penicillin. Very few patients received other antibiotic treatment for their EM. Of the 28 patients in the “other” group, eight patients received clindamycin, ciprofloxacin or cephalexin, one patient received chloramphenicol and the remaining 19 patients received mecillinam or trimethoprim. Most of these prescriptions were probably for another coincidental infection, such as urinary tract infections or conjunctivitis. The 9.2% of patients with EMs in whom we could not find a primary prescription appear to have been taken care of elsewhere to a large degree, which indicates that antibiotics are almost always prescribed for EMs in Norway. We have not been able to register the duration of antibiotic treatment in this study. For EM, antibiotic guidelines recommend 14 days. Although prophylactic antibiotic treatment after TBs is not recommended, almost 20% of patients are prescribed antibiotics. The distribution of antibiotics given for TB resembles the pattern of prescription for EM and it appears as though doctors prescribe for TB “as if it was an EM” (Figs. 3b and c). There was a higher degree of prescription at the OOHs for both EM and TB patients. This could reflect the difficulty of follow-ups. Another explanation may be that not all doctors working in the OOHs are GPs and are therefore not used to these diagnoses; hence, they may be more prone to “just in case” prescriptions.

4.5. Antibody testing The rate of antibody testing is high. Guidelines for EM treatment state that antibody testing is not recommended because EM is a clinical diagnosis. Less than 50% will have positive antibodies at the time of EM and the condition should be treated when diagnosed, regardless of antibody level (Hansen and Asbrink, 1989; Dessau et al., 2010; Stanek et al., 2012). However, there are some clinical indications to perform antibody testing at the time of EM, such as testing for systemic LB, but this is unlikely to explain the rate of about 20% among the RGPs. It has been shown that 9.6–18.4% of healthy blood donors are Borrelia IgG positive in endemic areas in Norway (Mygland et al., 2006; Hjetland et al., 2014). Given the high degree of seropositivity in the population, overuse of antibody testing can lead to false positive Lyme borreliosis diagnoses?. The low test rate in the OOHs is probably because of a lack of facilities for blood sampling and referral of patients to their RGP during business hours. The test rate for TBs only is almost identical. A simple TB is not an indication for Borrelia antibody testing, but in our study, we have not been able to register whether there are other good reasons for testing. It is tempting to believe that the high degree of antibody testing has more to do with the “just in case” behaviour of the doctor and/or the patient. It is also possible that the knowledge of tick-borne diseases is inferior to the fear of them, and leads to uncertainty even among doctors.

4.6. Referral In general, about 85% of consultations in general practice end there, regardless of diagnosis (Ringberg et al., 2013). Here, the very low number of referrals confirms that the first stage of LB in Norway is mostly seen and treated in primary care. The overlap between the GPs and OOHs of less than 4% also shows that EMs and TBs are quite easily treated in one single consultation. 4.7. Diagnoses given For EMs and TBs, we have only found relevant diagnoses in the EMRs in about 50% of cases. The diagnoses are more accurate at the OOHs. One explanation may be that, as more patients are seen there for the first time, the physician is obliged to enter a new diagnosis into the EMR to be able to get reimbursement for the consultation. At the RGPs, the diagnoses from the last consultations can easily be reused in the EMRs. There is no connection between diagnoses given, prescribing and reimbursement. It is well known that incorrect diagnosis coding in primary care is common (Botsis et al., 2010). With our method, we were only able to register whether a consultation was due to a TB and/or an EM, and not whether more serious infections or symptoms were the case. However, the “caution” proportion of the diagnoses, in which systemic LB and other tick-borne disease can “hide”, is only about 2%. Almost 25% of the diagnoses were found to be irrelevant. This suggests that using our method made it possible to find more actual TB and EM consultations than we would have done by other methods, which makes the incidence figures more reliable. 4.8. Strengths and limitations We have only been able to use consultation data from GPs with known list lengths, and have therefore missed some of the relevant consultations at their practices that were performed by other colleagues. Not being allowed to register the patients’ social security numbers forced us to find the overlaps between the RGP and the OOH cohorts through a text search. By excluding all possible overlaps, our figures are therefore the minimum figures. The figures are from 2005 to 2009, with a mean LB incidence of 6.5 LB/100,000 inhabitants/year. The LB incidence in Norway has shown a small increase in the years 2005–2015, ranging from 5.7 to 8.1 LB/100,000 inhabitants/year. The quality of the incidence estimate depends on getting representative data. The computer program was not compatible with all EMRs, which lowered the participation rate, but we obtained representative data from more than 50% of all the doctors who were able to participate. In addition, knowing each GP’s list length and each OOH’s catchment area enabled us to calculate the incidence from the available data. Had we instead divided the cases found by the whole population, the figures would have been falsely smaller. It is a strength of our study that we have also counted the OOH patients. OOH patients comprise about one third of the total cases and contribute to about one fifth of the incidence. Bypassing the diagnosis codes given and instead identifying TB and EM




8

consultations through the full text search has shown to be a good method of getting a more accurate estimate of the incidence. 5. Conclusion To estimate the Norwegian incidence of consultations for solitary EM, the skin lesion associated with early stage LB, we extracted data from the EMRs of RGPs and OOHs located in the four counties with the highest rates of registered cases of disseminated LB, in the 5 years from 2005 to 2009. We found an overall EM incidence of 448 EM/100,000 inhabitants/year in these counties, which yields a national incidence of 148 EM/100,000 inhabitants/year. In age groups younger than 40 years, most EM patients are male. In age groups older than 40 years, most patients are female. Overall, we found that older females have the highest rates of EM. The study also shows a high number of patients sought care for TBs without signs of infection, which indicates that the fear of tick-borne infections is high in the general population. PcV is the most commonly used antibiotic to treat EM, which complies with the national guidelines for antibiotic use. There is an overuse of antibiotics in patients with TBs without EM. Antibody tests were performed in 15% of the cases. Less than 1% of the patients were referred to secondary care. Acknowledgements We would like to thank all the participating GP offices and primary care out-of-hours services. We thank our GP colleague, Mark Fagan, for his thorough read-through of the manuscript. Funding for the study was granted by The University of Oslo, Institute of Health and Society and the Norwegian Research Fund for General Practice. Additional funding was received from the Antibiotic Centre for Primary Care, the NORM surveillance program for antibiotic resistance in human pathogens, the National Centre of Rural Medicine and the Eckbo Trust. None of those providing economic support were involved in the study design, data collection, analysis and interpretation of data, writing of the report nor in the decision to submit the article for publication. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ttbdis.2016.06. 006. References Bennet, L., Halling, A., Berglund, J., 2006. Increased incidence of Lyme borreliosis in southern Sweden following mild winters and during warm, humid summers. Eur. J. Clin. Microbiol. Infect. Dis. 25 (7), 426–432. Bennet, L., Stjernberg, L., Berglund, J., 2007. Effect of gender on clinical and epidemiologic features of Lyme borreliosis. Vector Borne Zoonotic Dis. 7 (1), 34–41. Berglund, J., Eitrem, R., Ornstein, K., Lindberg, A., Ringer, A., Elmrud, H., Carlsson, M., Runehagen, A., Svanborg, C., Norrby, R., 1995. An epidemiologic study of Lyme disease in southern Sweden. N. Engl. J. Med. 333 (20), 1319–1327. Botsis, T., Bassoe, C.-F., Hartvigsen, G., 2010. Sixteen years of ICPC use in Norwegian primary care: looking through the facts. BMC Med. Inform. Decis. Mak. 10, 11, http://dx.doi.org/10.1186/1472-6947-10-11. Dessau, R.B., Bangsborg, J.M., Ejlertsen, T., Skarphedinsson, S., Schonheyder, H.C., 2010. Utilization of serology for the diagnosis of suspected Lyme borreliosis in Denmark: survey of patients seen in general practice. BMC Infect. Dis. 10, 317. Hansen, K., Asbrink, E., 1989. Serodiagnosis of erythema migrans and acrodermatitis chronica atrophicans by the Borrelia burgdorferi flagellum enzyme-linked immunosorbent assay. J. Clin. Microbiol. 27 (3), 545–551. Hjetland, R., Nilsen, R.M., Grude, N., Ulvestad, E., 2014. Seroprevalence of antibodies to Borrelia burgdorferi sensu lato in healthy adults from western Norway: risk factors and methodological aspects. Apmis 122 (11), 1114–1124. Hofhuis, A., Harms, M., Bennema, S., van den Wijngaard, C.C., van Pelt, W., 2015. Physician reported incidence of early and late Lyme borreliosis. Parasite Vectors 8, 161.

Hubalek, Z., 2009. Epidemiology of lyme borreliosis. Curr. Probl. Dermatol. 37, 31–50. Lindgren, E., Jaenson, T., 2006. Lyme Borreliosis in Europe: Influences of Climate and Climate Change, Epidemiology, Ecology and Adaptation Measures. WHO Regional Office for Europe, Available from: http://www.euro.who.int/data/assets/pdf file/0006/96819/E89522.pdf (accessed 16.05.12.). Ljostad, U., Mygland, A., Skarpaas, T., 2003. Neuroborreliosis in Vest-Agder. Tidsskr. Nor. Laegeforen. 123 (5), 610–613. Morken, T., Hansen, E., Hunskar, S., 2009. Legevaktorganisering I Norge Rapport Fra Nasjonalt Legevaktregister 2009. National Centre for Emergency Primary Health Care, Available from: https://bora.uib.no/handle/1956/6248(accessed 16.05.12). Mygland, A., Skarpaas, T., Ljostad, U., 2006. Chronic polyneuropathy and Lyme disease. Eur. J. Neurol. 13 (11), 1213–1215. Nadelman, R.B., Nowakowski, J., Fish, D., Falco, R.C., Freeman, K., McKenna, D., Welch, P., Marcus, R., Aguero-Rosenfeld, M.E., Dennis, D.T., Wormser, G.P., 2001. Prophylaxis with single-dose doxycycline for the prevention of Lyme disease after an Ixodes scapularis tick bite. N. Engl. J. Med. 345 (2), 79–84. Nadelman, R.B., 2015. Erythema migrans. Infect. Dis. Clin. North Am. 29 (2), 211–239. Nygard, K., Brantsaeter, A.B., Mehl, R., 2005. Disseminated and chronic lyme borreliosis in Norway, 1995–2004. Euro Surveill. 10 (October 10), 235–238. O’Connell, S., 1996. European concerted action on Lyme borreliosis (EUCALB). Euro Surveill. 1 (3), 23–24. Ringberg, U., Fleten, N., Deraas, T.S., Hasvold, T., Forde, O., 2013. High referral rates to secondary care by general practitioners in Norway are associated with GPs’ gender and specialist qualifications in family medicine, a study of 4350 consultations. BMC Health Serv. Res. 13, 147. Rizzoli, A., Hauffe, H., Carpi, G., Vourc, H.G., Neteler, M., Rosa, R., 2011. Lyme borreliosis in europe. Euro Surveill. 16 (27). Smith, R., Takkinen, J., 2006. Lyme borreliosis: Europe-wide coordinated surveillance and action needed? Euro Surveill. 11 (6) (E060622.060621). Stanek, G., Fingerle, V., Hunfeld, K.P., Jaulhac, B., Kaiser, R., Krause, A., Kristoferitsch, W., O’Connell, S., Ornstein, K., Strle, F., Gray, J., 2011. Lyme borreliosis: clinical case definitions for diagnosis and management in Europe. Clin. Microbiol. Infect. 17 (1), 69–79. Stanek, G., Wormser, G.P., Gray, J., Strle, F., 2012. Lyme borreliosis. Lancet 379 (9814), 461–473. Stricker, R.B., Johnson, L., 2014. Lyme disease: call for a Manhattan Project to combat the epidemic. PLoS Pathog. 10 (1), e1003796. Tveitnes, D., Natas, O.B., Skadberg, O., Oymar, K., 2012. Lyme meningitis, the major cause of childhood meningitis in an endemic area: a population based study. Arch. Dis. Child. 97 (3), 215–220. Vandenesch, A., Turbelin, C., Couturier, E., Arena, C., Jaulhac, B., Ferquel, E., Choumet, V., Saugeon, C., Coffinieres, E., Blanchon, T., Vaillant, V., Hanslik, T., 2014. Incidence and hospitalisation rates of Lyme borreliosis, France, 2004 to 2012. Euro Surveill. 19 (34). Centers for Disease Control and Prevention (CDC), 2011. Case definition for notifiable Lyme’s disease. https://wwwn.cdc.gov/nndss/conditions/lymedisease/case-definition/2011/ (accessed 2016.02.29.). Centers for Disease Control and Prevention (CDC), 2015. Lyme Disease Incidence Rates by State, 2004-2013. http://www.cdc.gov/lyme/stats/chartstables/ incidencebystate.html (accessed 16.02.29.). Classification Committee of the World Organization of Family Doctors (WICC), 2011. ICPC-2. http://www.ph3c.org/4daction/w3 CatVisu/en/icpc-2-contenton-line.html?wDocID=81 (accessed 16.02.29.). European Concerted Action on Lyme Borreliosis (EUCALB), 2015. www.eucalb.com (accessed 16.02.29.). Norwegian Antibiotic Centre for Primary Care, 2015. National guidelines for antibiotic use in primary care. (Last update: 2015.09.23). www. antibiotikaiallmennpraksis.no (Norwegian) English version: http://www. helsebiblioteket.no/retningslinjer/antibiotika/antibiotic-treatment-concise (accessed 16.02.29.). Norwegian Directorate of Health, 2013. Norwegian guidelines for antibiotic use i hospital care. (Last update: 2013.06.19). http://sites.helsedirektoratet.no/sites/ antibiotikabruk-i-sykehus/Sider/default.aspx (Norwegian) (accessed 16.02.29.). The Norwegian Health Economics Administration (HELFO), 2016. https://helfo.no/ english/about-helfo (accessed 2016.02.29.). Norwegian Surveillance System for Communicable Diseases (MSIS), 2010. Lyme borreliosis statistics. www.msis.no (accessed 16.02.29.). Public Health Agency of Sweden, 2013. Laboratoriediagnostik av borreliainfektion. http://www.folkhalsomyndigheten.se/pagefiles/13257/laboratoriediagnostikav-borreliainfektion-2013-101-28.pdf (Swedish) (accessed 2016.02.29.). Regional Ethics Committee, 2010. (accessed 2016.02.29.) https://helseforskning.etikkom.no/ikbViewer/page/forside? ikbLanguageCode=us. Statistics Norway, 2010. Table 07459: Population, by sex and one-year age groups. https://www.ssb.no/statistikkbanken/SelectVarVal/Define. asp?MainTable=NY3026&KortNavnWeb=folkemengde&PLanguage=1&checked=true (accessed 16.05.06.).
