Acute Q fever in Majorca island 2017–2022. An underestimated problem

Abstract

Introduction

Q fever (QF) is a zoonotic infection caused by Coxiella burnetii. Previous studies suggest that the incidence in the island of Majorca may be high. The objective was to know the epidemiological, clinical, diagnostic, and therapeutic characteristics of acute QF (AQF) on the island of Majorca during the years 2017–2022.

Methods

Retrospective analysis of a series of cases of AQF diagnosed in 3 out 4 hospitals of the Health Service of the island of Majorca. AQF was considered if a clinically compatible syndrome was present with one of the following serological criteria: confirmed AQF if IgG seroconversion phase II, very possible if only IgG ≥1/128 is available in addition to a positive IgM, and AQF possible if IgG≥ 1/512 with negative IgM or if IgM positive with negative IgG in a single serological determination.

Results

223 cases of AQF were diagnosed, of which 102 (45.7%) were confirmed, 84 (37.7%) very probable and 37 (16.6%) probable AQF. Prolonged febrile syndrome was the most frequent clinical diagnosis (107, 48.0%), followed by pneumonia with/without pleural effusion (49, 21.9%), acute hepatitis (38, 17.0%), pericarditis and/or myocarditis (6, 2.6%). Three patients developed endocarditis (one in the acute phase, two others during follow-up). The median number of cases per year was 34 (p25–p75: 31.7–40.25). AQF-related mortality was 1.8% (4 patients).

Conclusion

AQF is a preventable and little recognized zoonosis, causing significant morbidity and mortality in Majorca.

Resumen

Introducción

La fiebre Q (FQ) es una zoonosis de distribución universal causada por Coxiella burnetii. Estudios previos sugieren que la incidencia en Mallorca puede ser elevada. El objetivo fue conocer las características epidemiológicas, clínicas, diagnósticas y terapéuticas de la FQ aguda (FQA) en la isla de Mallorca durante los años 2017 a 2022.

Métodos

Análisis retrospectivo de una serie de casos de FQA diagnosticados en 3 hospitales del Servicio de Salud de la isla de Mallorca. Se consideró FQA si el cuadro cínico era compatible y cumplían alguno de los siguientes criterios serológicos: FQA confirmada si seroconversión de IgG fase II, muy probable si solo se dispone de una determinación de IgG ≥1/128 además de IgM positiva y FQA probable si IgG ≥1/512 con IgM negativa o si IgM positiva con IgG negativa en una única determinación serológica.

Resultados

Se diagnosticaron en el periodo del estudio 223 casos de FQ aguda, de los cuales 102 (45,7%) FQA confirmada, 84 (37,7%) FQA muy posible y 37 (16,6%) FQA posible. El síndrome febril prolongado fue el diagnóstico clínico más frecuente (107, 48,0%), seguido de la neumonía con/sin derrame pleural (49, 21,9%), hepatitis aguda (38, 17,0%), pericarditis y/o miocarditis (6, 2,6%). Tres pacientes desarrollaron endocarditis (una en la fase aguda, otros dos durante el seguimiento). La mediana de casos por año fue de 34 (p25-p75: 31,7–40,25). Cuatro pacientes (1,4%) fallecieron en relación a la FQA.

Conclusión

La FQ es una zoonosis prevenible y poco reconocida, causante de una significativa morbimortalidad en Mallorca.

Introduction

Q fever is a universally-distributed zoonosis caused by Coxiella burnetii. The reservoir of this bacterium is extensive and includes arthropods, in addition to wild and domestic mammals and birds, with humans being an accidental host.^1,2C. burnetii is an extremely resistant microorganism in the environment. It is found in animal urine, faeces, milk, placenta and amniotic fluid.³ The most common source of infection in humans is domestic ruminants, especially goats and sheep,⁴ through airborne transmission by inhalation of spores, but oral and human-to-human transmission via the transplacental route during birth has also occasionally been reported. Transdermal transmission via vectors such as ticks has been demonstrated in animals but not in humans.⁴ Infection in animals is mostly subclinical and is associated with infertility or late abortions.³

In humans, the incubation period is estimated to be between seven and 32 days, with a median of 18 days.⁵ The spectrum of manifestations of the acute form (AQF) is variable, ranging from asymptomatic seroconversion (in 50-60%) to self-limiting febrile illness, or more severe manifestations, such as pneumonia, which can be critical, or acute hepatitis^1,2; other clinical forms such as meningoencephalitis⁶ and myopericarditis⁷ have also been reported. It is estimated that 2–5% of symptomatic cases require hospital admission, and up to 20% in some series. The chronic form or localised persistent infection is more serious, with valvular heart disease and/or immunocompromised patients being more susceptible, and it often manifests as infected aneurysms, endocarditis or spondylodiscitis.⁸

Cases of Q fever have been reported on all continents except Antarctica. In Europe, an outbreak in 2007–2010 was described, with more than 4000 people affected in the Netherlands,⁹ but since 2016, Spain has been the country reporting the highest number of cases to the European Centre for Disease Prevention and Control (ECDC).¹⁰ In Spain, the Centro Nacional de Epidemiología [National Epidemiology Centre] has considered Q fever a notifiable disease (ND) since 2015 and recommends individualised and weekly notification of each diagnosed case to the epidemiological surveillance centre (RENAVE; Red Nacional de Vigilancia Epidemiológica [National Epidemiological Surveillance Network]).¹¹ However, the Balearic Islands have not yet been incorporated into the national surveillance of Q fever as ND, as stipulated in Spanish Royal Decree (RD) 2210/1995. In 2022, 440 cases of C. burnetii infection were recorded in Spain, and 519 in 2023.¹² Its prevalence was highest in the autonomous region of the Canary Islands. Case series have been published in practically all parts of Spain,^13–19 including Majorca.²⁰ A Spanish study reviewing hospital discharge diagnoses coded as Q fever between 1998 and 2015 estimated an incidence rate in that period of 0.53 cases/100,000 person-years, with the highest incidence recorded in the Canary Islands and the Balearic Islands (1.48 and 1.43 cases/100,000 person-years, respectively), and an overall estimated mortality rate of 1.5/100 cases.²¹

Our aim with this study was to increase our understanding of Q fever in Majorca by studying the clinical, epidemiological, diagnostic and therapeutic characteristics of the cases diagnosed at three of the four Balearic Islands Health Service (IB-Salut) hospitals over a six-year period.

Methods

This was a retrospective descriptive study in three hospitals on Majorca, which provide hospital care to 41 of the island’s 48 basic health areas (85.4% of the population). The study included the period from January 2017 to December 2022 at the hospitals of Manacor, Son Llàtzer and Son Espases, the latter being the reference hospital in the Balearic Islands.

In all cases the clinical signs and symptoms were consistent and there was no alternative diagnosis. Detection of C. burnetii phase II IgG was performed in all samples by indirect immunofluorescence assay (IFA) with Coxiella burnetii IFA IgG reagent (Vircell), while detection of IgM was performed by IFA (Coxiella burnetii IFA IgM, Vircell) or chemiluminescence (Coxiella burnetii VirClia IgM Monotest, Vircell) according to the protocol of the diagnostic laboratory. Sera with phase II IgG ≥1/512, or when requested by the clinician for patient follow-up, were also tested for phases I and II IgG and IgM with the Focus Coxiella phases reagent (Focus Diagnostics® Cypress, CA, USA). The serological criteria for diagnosis of the different conditions were based on the Centro Nacional de Epidemiología [National Epidemiology Centre] criteria¹¹ with a modification to increase specificity in cases of probable AQF: the IgM result was considered positive, and in cases with a negative result, the titre considered positive for IgG was increased. The established criteria were therefore: 1) "confirmed AQF" when there was evidence of seroconversion of IgG titres to C. burnetii within 4–8 weeks; 2) in cases where a single serological determination was performed, "highly probable AQF" was defined if IgG was ≥1/128 with IgM also positive; and 3) a third group, "probable AQF", was considered when IgG serology was positive at a high titre (IgG ≥1/512) with negative IgM or when IgM was positive with negative IgG. For the analysis, we collected demographic and epidemiological data, risk factors (structural heart disease, HIV infection or other immunosuppression), clinical, analytical and radiological data, the need for hospital admission, treatment received and complications from the patient's medical records. An anonymised database was created in which the confidentiality of the data was guaranteed for the storage and analysis of the variables using the statistical package SPSS 12.0 for Windows. Descriptive statistical analysis was performed by calculating frequencies for each of the qualitative variables and the median and percentiles [p25–p75] for the quantitative variables. For comparison of qualitative variables, the Chi-square test with Yates correction was used where necessary, considering a difference to be statistically significant if p < 0.05. Incidence rates were calculated by dividing the number of cases per year/population in the health areas corresponding to the three hospitals in that year according to information from the Spanish Instituto Nacional de Estadística [National Institute of Statistics] and expressed as cases/100,000 person-years.

The study was approved by the Hospital Son Espases Research Committee (No. CI-562-21, and in its subsequent extension No. CI-922-24). In accordance with the applicable provisions of current regulations and the ethical principles for medical research on human subjects, Informed Consent was not requested from the patients as this was an anonymised retrospective study.

Results

A total of 223 cases of AQF were identified, 102 (45.7%) of which were confirmed AQF, 84 (37.7%) highly probable AQF and 37 (16.6%) probable AQF. The median age of the cases was 49 years (p25–p75: 39–62), and 158 (70.9%) of the 223 cases were male.

The number of cases diagnosed each year (total and per hospital) is shown in Fig. 1. There was a median of 34 cases per year (p25–p75: 31.7–40.25). There was a larger number of cases in 2018. The estimated annual incidence in the health areas covered by the three hospitals is shown in Fig. 1b.

Figure 1.

a. Cases of AQF diagnosed each year at each of the hospitals in the study. b. Estimated incidence of cases per year in the health area of the study (cases/100,000 person-years).

Table 1 shows the epidemiological characteristics of all the cases. The majority of cases occurred in spring and winter, but there was also a considerable number in summer. A large number of patients lived in urban areas and had no direct contact with animals. In terms of the comorbidities analysed, 27 patients (12.1%) had some form of immunosuppression and seven (3.1%) had advanced chronic liver disease. In total, 21 patients (9.4%) had structural or valvular heart disease, and six of them had a prosthetic valve. Three patients (1.3%) had aortic aneurysm and four patients (1.8%) had a prosthetic joint. The median number of days of symptoms prior to consultation/admission was seven (p25–p75: 3–10). A total of 184 cases (82.5%) required hospital admission, with a median length of stay of six days (p25–p75: 3–10); 74.4% of patients were admitted to general internal medicine and infectious diseases departments, and six required direct admission from Accident and Emergency to intensive care units. Prolonged febrile syndrome was the most common clinical diagnosis (109, 48.9%), followed by pneumonia with/without pleural effusion (50, 22.5%) and acute hepatitis (37, 16.6%). Acute hepatitis was significantly more common in the group of patients with confirmed AQF, with 29 cases (28.4%) vs 8 cases (9.5%) in the highly probable/probable AQF group (p < 0.001), and pneumonia was significantly more common in the highly probable/probable AQF group, with 40 cases (17.9%) vs 10 (4.5%) (p < 0.001). No other differences were detected between the three groups of confirmed, highly probable or probable AQF. Other clinical diagnoses were pericarditis/myocarditis, meningoencephalitis, lymphadenitis, pleural effusion, arthritis, exanthema and oral aphthae. One female patient was diagnosed with mitral endocarditis without microbiological isolation, with a serology consistent with AQF, but it was not possible to perform phase I and II serology (Table 2).

Table 1.

Epidemiological data from the series of 223 patients diagnosed with AQF.

		n	%
Gender	Male	158	70.9%
	Female	65	29.1%
Season	Spring	74	33.2%
	Summer	58	26.0%
	Autumn	25	11.2%
	Winter	66	29.6%
Work and residential environment and contact with rural areas	Urban	106	47.5%
	Rural	73	32.7%
	Livestock farmer, veterinarian, agricultural worker	10	4.5%
	Recreational	3	1.3%
	Unknown	31	13.9%
Travel	Yes	14	6.3%
	No	148	66.4%
	Unknown	61	27.4%
Contact with animals	No	38	17.0%
	Goats	2	0.9%
	Sheep	6	2.7%
	Cows	0	0.0%
	Dogs/cats	102	45.7%
	Farm animals	11	4.9%
	Tick	1	0.4%
	Unknown	63	28.3%

Table 2.

Symptoms and main clinical diagnosis of the 223 cases of AQF.

		n	%
Symptoms	Fever	179	80.6
	Cough	68	30.6
	Expectoration	24	10.8
	Dyspnoea	41	18.5
	Headache	73	32.9
	Arthromyalgia	81	36.5
	Skin rash	8	3.6
	Gastrointestinal symptoms	33	30.0
	Meningism	6	2.7
	Profuse sweating	39	17.6
	Hepatomegaly	17	7.7
	Splenomegaly	15	6.8
Main syndrome diagnosis	Febrile syndrome	107	48.0%
	Pneumonia ± pleural effusion	49	21.9%
	Acute hepatitis	38	17.0%
	Pericarditis/myocarditis	6	2.7%
	Asthenia	5	2.2%
	Lymphadenitis	4	1.8%
	Aphthae/arthralgia/exanthema	3	1.3%
	Splenomegaly/abnormal liver function	2	0.9%
	Meningoencephalitis	1	0.4%
	Other	7	3.1%

Abnormalities in laboratory test variables are shown in Table 3. Liver function tests were considered abnormal when values exceeded twice the upper limit of normal, which was found in 112 (50.2%) of the total number of patients diagnosed with AQF. All patients diagnosed with hepatitis had significantly abnormal liver function tests, as did 58 (54.2%) of those with prolonged febrile syndrome and 10 (20.8%) of those diagnosed with pneumonia. Anticardiolipin antibodies were not measured in any of the cases.

Table 3.

Laboratory test abnormalities detected in AQF cases (confirmed and highly probable/probable AQF).

	n	%
Leukocytosis (leukocytes >11,500/mL) N = 219	46	21.0%
Anaemia (haemoglobin <12 g/dl) N = 206	38	18.4%
Thrombocytopaenia (platelets <120,000/mL) N = 218	62	28.4%
Increased aspartate aminotransferase (AST > 68 U/l) N = 193	75	38.9%
Increased alanine transaminase (ALT > 110 U/l) N = 210	71	33.8%
Increased gamma-glutamyl transferase (GGT > 128 U/l) N = 210	82	39.0%

In terms of imaging tests, chest X-rays were performed in 206 patients (92.4%) and abnormalities were found in 67 (32.5%). The most common radiological findings were pulmonary infiltrates in 46 cases and pleural effusion in nine cases. Transthoracic echocardiograms were performed in 43 patients (19.3%) at diagnosis or in the first few months of follow-up of the AQF episode (completed with transoesophageal echocardiogram in 10 cases), showing pericardial effusion in three cases, native valve endocarditis in two (in one case doubtful) and prosthetic valve endocarditis in another. Echocardiogram was performed if the patient had a risk factor for chronification (immunosuppression, prosthetic valve or endovascular prosthesis) or at the discretion of the treating physician. Abdominal imaging (ultrasound or computerised axial tomography [CT]) was performed in 98 cases (43.9%), and abnormalities were found in 42, most commonly hepatosplenomegaly (20 cases). Positron emission tomography (PET-CT) was performed in 13 cases, which was normal in all but one case, where a concomitant lung neoplasm was diagnosed.

At least 181 patients (77.7%) received some antibiotic therapy, the majority doxycycline (134 [60%]) and quinolones in 32 (14.3%). The mean length of antibiotic therapy was 12.5 days (SD: 5.05). Serological follow-up at six months was performed in 106/223 patients; 22 patients had elevated phase I antibody titres (>1/800), with one developing prosthetic aortic valve infective endocarditis, another possible native aortic valve endocarditis, and one persistent liver enzyme abnormalities. All three patients were treated with doxycycline and hydroxychloroquine. The remaining 19 patients were symptom-free; four of them were treated with doxycycline and hydroxychloroquine. Four patients died, three of them while in hospital: two from critical pneumonia and one from multiple-organ failure related to acute endocarditis. One patient died six months after the AQF episode due to prosthetic valve endocarditis.

Discussion

This study was carried out over a six-year period in three of the four public acute care hospitals on the island of Majorca which provide hospital care to 85.4% of the population. This allowed us to obtain a good perspective of the epidemiological situation and to estimate the incidence of acute infection. However, due to the retrospective nature of the study, this could be underestimated, as we only included cases diagnosed in public hospitals and not in primary care, and we were unable to include those diagnosed in the private medical sector. We followed strict diagnostic criteria for probable AQF by considering IgG titres ≥1/512, avoiding false positives, as we found persistent low titres in the asymptomatic population, suggesting a high seroprevalence, in line with reports from other regions.²² All of the above means we are able to affirm, as already suggested,^20,21 the importance of this zoonosis in Spain. The number of cases varies from year to year, peaking at an incidence of seven cases/100,000 person-years in 2018, but we do not have the means to study whether or not this was an outbreak. The incidence in 2022 of five cases/100,000 population is much higher than that reported in the same year in other Autonomous Regions in Spain. In addition, cases were also reported in 2020 (25 cases between April and December 2020), when the wearing of masks to protect against airborne transmission was mandatory even in open spaces, suggesting that there may be transmission mechanisms other than airborne, or perhaps transmission by domestic animals. As in other series¹³ and in the RENAVE registry,¹² the patients in our series were mostly middle-aged males. However, we found that they very often did not live in rural areas or have direct contact with livestock, as has been classically described.²³ The highly resistant C. burnetii spores can travel up to 18 km on winds, with the highest infectivity rate considered to be within a radius of 5 km²⁴; on Majorca, sheep are found across the island and frequently graze in the vicinity of towns and villages. The preliminary report of a recent study carried out on 165 small ruminant livestock farms, mostly with sheep, on the island of Majorca shows C. burnetii DNA to be present in 86.1% of the environmental samples.²⁵ Environmental determinants such as wind, humidity and temperature could all affect the transmission of infection. Other animals, such as dogs and cats, and even pigeons, rodents and other wild animals²⁶ can be carriers of C. burnetii. In our series, contact with dogs and cats was very common, and two patients recalled tick bites. Other alternative routes of transmission are still poorly understood, such as oral transmission through consumption of unpasteurised milk or dairy products or bites from ectoparasites such as ticks.⁴

There is great variability in the clinical presentation of AQF, which has been related to factors such as host susceptibility, inoculum size, route of infection and also geographical location.²⁷ In our series of AQF cases, we found that prolonged febrile syndrome was the most common form of presentation, followed by hepatitis and pneumonia. These data differ from the series previously described in the Balearic Islands,²⁰ in which pneumonia was the most common form of involvement, as has been reported in the north of the Iberian Peninsula,^15,19 in Catalonia¹⁸ and also in Madrid.¹⁴ Prolonged fever was described as the most common symptom of AQF in Andalusia,^16,21 associated with abnormal liver function tests. Endocarditis has very rarely been reported in AQF²⁸ and could be related to an antiphospholipid syndrome (Libman-Sacks endocarditis),²⁹ but in our case no anticardiolipin antibody test was performed to confirm this.

The six-month serological follow-up of patients diagnosed with AQF was not done routinely but at the discretion of the treating physician. There is debate in the literature regarding follow-up after AQF. Some authors suggest six-month follow-up serology for all patients after AQF,³⁰ while others only for those with risk factors for chronification.² The phase I antibody titre, which should be considered suggestive of QF complications and which should prompt further investigations, is also not well defined; 1/800 or 1/1,024.^2,31 There is no agreement on whether or not to perform an echocardiogram and test for anticardiolipin antibodies in all cases of AQF, with Dr Raoult²⁹ arguing they should be done, but others recommending echocardiogram only in cases with cardiovascular risk factors and in pregnancy.² A recent study conducted at the Hospital de Valme in Seville has convincingly demonstrated the limited utility of serological follow-up for at least six months in cases of AQF for the diagnosis of persistent infection; 8.1% of patients diagnosed with AQF had phase I IgG titres >1/1024, but they were asymptomatic and no persistent infection was detected in any of the cases.³²

Our study has important limitations, mainly due to its retrospective nature. Epidemiological data and serological follow-up were not available for all patients. In addition, it is possible that episodes of AQF may have gone unnoticed, especially in cases of pneumonia, but also in cases of acute hepatitis, as serological testing for QF is not routinely performed in these clinical situations, but is performed in cases of prolonged fever without a source, and this may have led to a bias in clinical diagnoses and an underestimation of the number of cases. In more than half of the cases, the diagnosis is probable AQF, as a second serological study was not performed after 2–3 weeks.

In conclusion, Q fever is a common disease on Majorca and the morbidity and mortality rates are considerable. Cases have been observed at all times of the year, in patients without the characteristic epidemiological factors of the disease, and even at the time of the COVID-19 pandemic, when the use of face masks in open spaces was mandatory. Under-diagnosis is also very likely, as serological studies for C. burnetii are not requested in all cases of pneumonia or acute hepatitis. We consider it necessary to establish a uniform QF surveillance system in each of Spain's Autonomous Regions, both in humans and in domestic animals and livestock, in order to achieve better characterisation of the clinical picture in our setting and its transmission mechanisms and reservoirs, and thus be able to introduce measures to prevent transmission.