Clinical features of 2041 human brucellosis cases in China

Background Human brucellosis has become a major public health problem in China. However, the available clinical data on brucellosis cases are limited. Methods We retrospectively reviewed medical charts of 2041 patients with confirmed human brucellosis and prospectively recorded their outcomes by telephone interview. These patients were admitted to the Sixth People’s Hospital of the Xinjiang Uygur Autonomous region between 1st January and 31st December 2014. Data on these patients were collected from hospital medical records. Results Many patients presented with fatigue (67%), fever (64%), arthralgia (63%) and sweating (54%). High erythrocyte sedimentation rate (ESR) (69%), high C-reactive protein (CRP) (39%), high alanine aminotransferase (ALT) (33%) and high aspartate aminotransferase (AST) (20%) were the most common laboratory findings, especially in acute patients. There was evidence of focal involvement in 90% of patients. A total of 61.5% of brucellosis patients recovered. Multivariate logistic regression analyses suggested that the risk factors key to unfavorable prognosis were: age≥45 years (OR = 1.75, 95% CI 1.36–2.24), back pain (OR = 1.50, 95% CI 1.16–1.94) and joint tenderness (OR = 1.73, 95% CI 1.13–2.65). The increasing duration of the illness increased the risk of poor prognosis. Conclusion Patients with brucellosis showed different characteristics in different clinical stages. In China, the chronicity rate of human brucellosis is high. The risk of poor prognosis is increased in patients aged 45 years or older, patients who have had brucellosis for a substantial period of time, and patients with back pain or joint tenderness. The clinical management of brucellosis should be improved to include sensitive diagnostic methods for subacute and chronic brucellosis.

Introduction Brucellosis, caused by bacteria of the genus Brucella, is a major zoonosis worldwide. Although it has been well controlled in most developed countries, brucellosis is still endemic in Africa, Asia, the Middle East, the Mediterranean and South America [1,2]. Human brucellosis is a major public health problem that has re-emerged in China since the mid-1990s, with the highest recorded number of cases (56,989 cases) in 2015 [3][4][5]. This may be due to the dynamic growth of animal husbandry in China, which enhances the chance of human infection. Moreover, the national infectious disease surveillance system has reported an increased rate of human brucellosis [6].
The disease is spread to humans mainly by the ingestion of infected meat or unpasteurized dairy products, by contact with infected animals or inhalation of infectious aerosolized particles [7]. Brucellosis usually leads to protean manifestation and may have polymorphic features affecting any organ system. The diagnosis of brucellosis is challenging because unusual presentations and non-specific symptoms can lead to misdiagnosis and treatment delay. Although the mortality rate is low, brucellosis can be severely debilitating and disabling [8,9]. A timely and accurate diagnosis is key to the clinical management of brucellosis. There is an urgent need for an increased awareness of the clinical characteristics of brucellosis among physicians.
Between 2000 and 2012, research into the clinical characteristics of brucellosis was done in Turkey [10], Iran [11], Greece [12] and Saudi Arabia [13]. But the results differed between studies. This may be due to small sample sizes or different methods for case classification. In order to reliably examine the clinical characteristics of brucellosis, the series used in the present study is larger than that used in previous studies, and a common case classification is used.
Reliable data on the clinical characteristics of human brucellosis should improve the identification of relevant prognostic factors and improve the clinical management of this debilitating condition. This study aims to describe the symptoms, signs and laboratory findings of 2041 hospitalized patients with laboratory confirmed brucellosis, and assess the risk factors for their clinical outcomes.

Setting
The Sixth People's Hospital of the Xinjiang Uygur Autonomous Region (hereafter called the Sixth People's Hospital) is a tertiary infectious disease specialist hospital with 11 clinical departments and 460 beds. As an important referral center, the hospital approximately reported 27% of total reported brucellosis cases in Xinjiang yearly. A total of 2041 patients with laboratory-confirmed brucellosis were admitted to the hospital between January 1 st and December 31 st of 2014 and were included in this retrospective study. We accessed patient medical records from March 12 th to early April 2016 and performed telephone interviews in late April 2016. The patient files were examined using a standardized form, which recorded demographic data, medical history, clinical and laboratory findings, any antibiotic treatment, and any focal involvements.

Case definition
A confirmed case of brucellosis was defined as a patient with compatible clinical symptoms (such as arthralgia, fever, sweating, chills, headache, myalgia, malaise) and laboratory evidence of Brucella infection diagnosed by bacteria culture or SAT (�1:200). Based on the duration of the systemic disease before admission to hospital, patients were divided into three groups: acute brucellosis (<2 months), subacute brucellosis (2-12 months), and chronic brucellosis (>12 months). All cases underwent routine laboratory tests.

Treatment and follow-up
Patients were treated with various combinations of antibiotics. The antibiotics regimens were given on the basis of China's Practice Guideline for brucellosis diagnosis and treatment [14]: doxycycline (100mg every 12 h), rifampin (600 -900mg every 24 h), intramuscular streptomycin (15mg/kg every 24 h), levofloxacin (200mg every 12 h), ciprofloxacin (750mg every 12 h), and co-trimoxazole (960mg every 12 h). For children under 8 years of age and pregnant women, only rifampin (600-900mg every 24 h) was provided. Given the effectiveness of previous treatment and patient compliance (Xinjiang's vast land posed a challenge to subsequent patient visits), the treatment duration was extended to 3-6 months for all patients. The outcome of treatment (including medication compliance and symptom recovery) was investigated by telephone interview. Admittedly telephone interview is not the ideal method of follow-up, but since in chronic/ persisting brucellosis the symptoms are often subjective and telephone interview allowed for a larger enrollment of patients in follow-up, we considered this to be an acceptable follow-up method.

Data analysis
Data were entered into Epidata (version 2.0, Odense, Denmark). Medians and interquartile ranges (IQRs) were calculated for continuous variables, and compared between different groups using the Wilcoxon rank sum test. For categorical variables, case frequencies in each category were compared using a Chi-square test or Fisher's exact test. For multiple comparisons, the Bonferroni correction was applied. To further examine the association between potential risk factors and treatment outcome, we first performed univariate analysis. For multivariable logistic regression, we included variables with p<0.05 in univariate analysis or those believed to be potential risk factors associated with the outcome. The software program SAS 9.3 and R 3.3.1 were used to analyze the data.

Ethical Approval
Ethical approval (no. 201533) was obtained from the Institutional Review Board (IRB) of China CDC before the survey began. With IRB's ratification, we signed a confidentiality agreement with the hospital to use patient medical records for research purposes. We confirmed that all participant identifying information (including patient names, ID numbers, home addresses and telephone numbers) would not be included in recordings, written descriptions or publications. Because written informed consent is difficult to obtain during a telephone survey, the use of verbal informed consent was approved by the IRB for this retrospective study. Verbal informed consent was obtained from all respondents over the phone and documented in forms. For subjects younger than 17 years of age, we obtained verbal consent from their parents or legal guardians.

Laboratory findings
Laboratory findings at admission are shown in Table 3. The most common laboratory findings were high ESR (69%), high CRP (39%), high ALT (33%) and high AST (20%). These abnormal findings more frequently occurred in acute cases than in the other two groups, adjusted p<0.05 (S1 Fig). Anemia was more commonly seen in acute cases (26%) than in subacute cases (20%), adjusted p<0.05. Across all patients, the median hemoglobin was 129 g/L (IQR 117-139), median leukocyte count was 5.9×10 9 cells/L (IQR 4.7-7.1), median lymphocyte count was 2.8×10 9 cells/L (IQR 2.1-3.7) and median thrombocyte count was 221×10 9 /L (IQR 175-272). The median CRP and ESR in the total group were 6.6 mg/L (IQR 2.4-20.0) and 28 mm/h (15-46) respectively. A total of 903 of 1925 cases (47%) showed Brucella growth in their blood culture. And the positive rate (53%) in acute cases was significantly higher than in subacute (31%) and chronic groups (19%) (adjusted p<0.0001). The positive rate of SAT of acute and subacute cases gradually decreased along with the progress of the condition after treatment (S2 Fig). The results of regular re-examinations after discharge showed that most acute patients laboratory tests improved after treatment (S3 Table).

Follow up and outcomes
A total of 1321 cases (65% of total cases) responded to our telephone interview more than 450 days after discharge from hospital. This was long enough to observe treatment outcome (S2 Table).
Multivariate analysis. Multivariate logistic regression analysis was used to identify factors that were independently associated with poor prognosis. Variables with p<0.05 in univariate analysis were included in multivariate analysis (Table 6). We found that fever, headache, arthralgia, aminotransferase elevation, culture positive and initial treatment were no longer significantly (p>0.05) associated with unfavourable prognosis in multivariate analysis.

Discussion
This study retrospectively collected data from the medical records of 2041 patients with laboratory confirmed brucellosis and followed up their clinical outcomes. The clinical features of human brucellosis were described and risk factors for unfavorable prognosis were assessed systematically. The results from this study provide a theoretical basis for clinical diagnosis, treatment and case management of human brucellosis.
The gender and age distributions were consistent with previous studies [5,13]. A total of 87% of children under 15 years of age presented as acute cases, which suggests that children seek medical advice earlier than adults, or they have more symptoms with infection than adults and are therefore more likely to present to hospital earlier. Or children may be less likely to have localized infections, which may not present as rapidly. The primary transmission route of brucellosis in our study was through occupational exposure (97%), which is in accord with the results of epidemiologic investigations in China [3,15]. In other endemic countries, infections occur mostly due to ingestion of unpasteurized dairy products [16,17]. In our study, the most common clinical manifestations of human brucellosis were fatigue (67%), fever (64%), arthralgia (63%) and sweating (54%), of which the rate was lower than that reported by the WHO [18]. We found that patients in different clinical stages presented with different clinical features. Systematic manifestation, such as fever, lack of appetite and weight loss, were more frequent in acute cases, while arthralgia, back pain and joint tenderness occurred more frequently in patients with longer disease durations. Similar findings have been described in a study done in Turkey [19].
Elevated liver enzymes, increased CRP and increased ESR were the most common laboratory findings seen in our series, especially in acute cases (see Table 3). Buzgan et al. reported similar results [19]. Along with increased illness duration, we found that both the positive rates of SAT and blood culture fell. This suggests that we should use tests with higher sensitivity to diagnose subacute and chronic cases to avoid missed diagnosis. ELISA, PCR, Coombs test and bone marrow culture may be good choices for subacute and chronic patients, especially for those who have used antibiotics [20,21].
In our study, osteoarticular involvement was observed in 64% of acute cases, 77% of subacute cases and 87% of chronic cases. This is higher than that reported by Buzgan et al. [19] and Hasanjani et al. [22]. A study done in Iran showed that sacroilitis (75.7%) was more common than spondylitis (21.4%) and peripheral arthritis (8.6%) [23], while a survey done in Turkey found that peripheral arthritis (56.5%) occurred more frequently than sacroilitis (24.6%) and spondylitis (12.3%) [19]. Our study indicated that peripheral joints (57%) were more likely to be affected by Brucella than spine (20%) and sacroiliac joints (2%). A total of 122 patients (6%) showed evidence of respiratory involvement, which mainly manifested as acute infection (92%) with no chronic cases. This is consistent with another report done in Turkey [23]. This suggests that respiratory involvement mainly occurred in acute cases. However, the 75% of cases with respiratory abnormalities presented as bronchitis in our series, while pneumonia (68.4%) and pleural effusion (30.8%) were more commonly found in Turkey [24].
In some studies from Iran, the recovery rate of human brucellosis has been reported between 80.6% [25] and 89.49% [26], which is higher than that found in our study (61.5%). In the Sixth People's Hospital, many referral patients were received. Therefore, a more severe disease population with a greater risk of poor prognosis may have been included in the current study. A small amount of research has been done on the factors that influence prognosis of human brucellosis. In our study, we identified five independent factors that influenced whether a patient with human brucellosis received a poor prognosis: age �45 years, back pain, joint tenderness, long duration of illness and haematologic involvement.
Brucellosis in older patients increases the risk of poor prognosis because these patients often have some immune system dysfunction [27][28][29]. Other studies, however, did not find differences in outcomes according to patient age [26,30]. In our series, patients aged �45 years were more likely to have unfavorable prognosis. We also found that back pain and joint tenderness were associated with increased risk of poor prognosis, which is consistent with Wu's finding [31].
Moreover, patients who had brucellosis over a longer duration were more likely to have unfavorable prognosis after treatment. Indeed, Alavi and colleagues reported that the majority of patients who relapse had a prolonged duration of time between the appearance of symptoms and initiation of treatment [32]. Although abnormal haematologic findings have previously been shown to be a marker of poor prognosis [30,33], we present the opposite result (OR = 0.60, 95% CI 0.45-0.79, p = 0.003). This may be because those who had abnormal haematologic findings were more likely to present as acute cases in our series, and acute cases tend to have a lower rate of poor outcomes. Therefore, this association could be spurious. More research is needed in this area.
Previous meta-analyses have reached different conclusions regarding the preferred regimens for brucellosis [34,35]. Generally, dual or triple regimens are advisable. In our series, multivariate analyses showed no significant differences between each combination therapy. However, since the data on real treatment duration and patient compliance after discharge were insufficient, this conclusion about the association between regimens and treatment outcome is substantially underpowered. This would need to be further investigated in a prospective study.
Our study was limited to available data of brucellosis cases identified through retrospective analysis. Cases included in this study were all from a provincial tertiary hospital, where many severe patients were referred. This might explain why the recovery rate observed in our series is lower than that identified in previous studies. Clinical management was uncontrolled, pathological data and medication compliance data were not available, and relapse and reinfection were hard to distinguish from treatment failure. Any differences in outcome cannot be interpreted to be due to the use of antibiotics since there are no clear diagnostic criteria for measuring real cure, and telephone follow-up may decrease the validity of the findings.
In summary, brucellosis cases have a high rate of chronicity. The longer the illness duration, the more difficult brucellosis is to cure. To improve the clinical management of brucellosis in China, early diagnosis and treatment should be of high priority. Age �45 years, back pain, and joint tenderness are risk factors of adverse prognosis of brucellosis. Longer courses of treatment should be considered for older patients and those with osteoarticular involvement. The positive rate of blood culture and SAT may gradually decrease with the progress of the condition. Therefore, sensitive diagnostic methods should be used to identify subacute and chronic brucellosis cases to avoid missed diagnosis. Brucellosis patients may show different characteristics in different clinical stages: abnormal laboratory findings and respiratory involvement may be more frequent in acute cases, osteoarticular involvement may be more frequent in subacute and chronic cases.