| Objectives:Here,we aimed to investigate the prevalence and etiology of high fever in hospitalized tuberculosis patients,and the effects(such as the proportion of death or transfer,rates of antibiotic use and the duration of hospital stay)on clinical experience.Methods:1.A six year retrospective study was conducted using the medical records of hospitalized patients with tuberculosis of Department of Infectious Disease in comprehensive hospital in Kunming between January 2010 and December 2015.Firstly,the prevalence and etiology of high fever were analysed in tuberculosis patients and clinical effects(such as the proportion of death or transfer,rates of antibiotic use and the duration of hospital stay)were compared among the high fever group(HF group),the low-and moderate-grade fever group(LM group)and the afebrile group(AF group).Secondly,in accordance with the matching criteria of the same first discharge diagnosis,admission year,gender,and similar age,a case-control study was conducted.Then,a subgroup analysis was conducted to compare the clinical characteristics in the high fever patients with and without other bacterial or fungal infections.Finally,in order to validate the results,a cohort of patients from Yuxi People's Hospital in Yuxi City and 2016 data from the Kunming General Hospital of Chengdu Military Command.2.Statistical analyses were conducted using IBM SPSS Statistics 20(IBM,Armonk,NY,USA)and the figures were drawn using the GraphPad Prism software package(version 5 for Windows;GraphPad Software,San Diego,CA).Categorical variables were expressed as n(%)and continuous parameters as mean ± standard deviation(Mean ± SD)or median(IQR).Pearson x2 test or Fisher's exact test were used to compare the categorical variables,while either the Student's T test or the Mann-Whitney U test was used for comparison of pairs of continuous variables,and either the one-way ANOVA or Kruskal-Wallis H test were used for comparisons of more than two means.All statistical analyses were based on a two-tailed hypothesis tests,and the criterion for statistical significance was set at P<0.05.Results:1.The prevalence of high fever in hospitalized tuberculosis patientsA total of 2166 inpatients with tuberculosis were enrolled in Kunming cohort,1362(62.88%)of them were male,1764(81.44%)of them were Han ethnicity.Sixty-nine patients were defined as exclusion group(EX group)due to either death or transfer from our facility before their body temperatures returned to normal(35 were with high fever,30 were with low-and moderate-grade fever,4 were afebrile).The incidence of fever in all tuberculosis patients was 42.84%(928/2166,95%CI 40.76%-44.92%),and the incidence of high fever was 12.00%(260/2166,95%CI 9.32%-14.68%).In the individual years between 2010 to 2015,the prevalence of fever and high fever in the tuberculosis patients ranged from 37.74%to 46.53%and from 9.74%to 14.76%,respectively.2.The etiology of high fever in tuberculosisThirty-fifth of 260 tuberculosis patients with high fever were either death or transfer from our facility before their body temperatures returned to normal,the remaining 225 patients with high fever divided into two groups based on whether a probable cause for high fever that was unrelated to the active tuberculosis infection was identified by the clinicians(n= 115,51.11%)or whether no cause unrelated to the tuberculosis infection was found(n=110,48.89%).The most common cause of high fever was other bacterial infection(62/225,27.56%),followed by fungal infection(28/225,12.44%),drug-induced fever(10/225,4,44%),upper respiratory infection(8/225,3.56%),Herxheimer-like reaction(4/225,1.78%)and post-operative fever(3/225,1.33%).3.Effects of high fever on patients with tuberculosisThese EX patients accounted for 13.46%(35/260,95%CI 9.31%-17.61%)of the total patients with high fever.This was 7.56 times higher than the proportion of EX patients in the group without high fever(1.78%,34/1906,95%C1 1.19%-2.37%;P<0.0001).The median durations of hospital stay in the HF,LM and AF groups were 23 days(IQR 15-33),19 days(IQR 13-27)and 16 days(IQR 11-23)respectively.The statistical significances of the differences between these groups(P<0.0001).The median duration of antibiotic use in the HF group was 10 days(IQR 0-17),compared to 0 days(0-9)in the LM group and 0 days(0-4)in the AF group(P<0.0001,P<0,0001)respectively.In the HF group,the proportion of patients who used antibiotics was 74.67%(168/225,95%CI 68.99%-80.35%),which was significantly higher than the proportion using antibiotics in the LM(49.06%,313/638,95%CI 45.18%-52.94%)and AF(28.77%,355/1234,95%CI 26.24%-31.30%)groups(P<0.0001,P<0.0001).A paired comparison analysis showed the same conclusion.Tuberculosis patients with both high fever and other bacterial or fungal infections were defined as HF1*group(n=70),tuberculosis patients with high fever but no other bacterial or fungal infections were defined as HF2*group(n=85).The duration of fever(11.50 ± 8.55 d vs.8.67 ± 8.07 d;P=0.0368),the duration of antibiotic use(17.54 ± 12.16 d vs.9.44 ± 11.82 d,P<0.0001),the frequency of antibiotic use(91.43%[64/70,95%CI 84.87%-97.99%]vs.63.53%[54/85,95%CI 53.30%-73.76%];P=0.0001),the WBC levels(10.25×109/L[IQR7.94-14.79]vs.6.88×109/L[(5.23-8.50];P<0.0001)and the ESR levels(70.11 ± 36.16 mm/h vs.56.36 ± 35.30 mm/h);P=0.0189)in HF1*group were all significantly higher than in HF2*group.4.Validation analysisThe incidence of fever in all tuberculosis patients was 56.94%(283/497,95%CI 52.59%-61.29%),and the incidence of high fever was 14.69%(73/497,95%CI 11.58%-17.80%)according to the datas in Department of Infectious Disease in the Yuxi People's Hospital from 2015 to 2016;and the incidence of fever in all tuberculosis patients was 37.85%(123/325,95%CI 32.58%-43.12%)the incidence of high fever was 10.15%(33/325,95%CI 6.87%-13.43%)according to the datas in Kunming from 2016.And those datas shared the similar clinical effects with Kunming test cohort.Conclusions:High fever has a high incidence in the hospitalized patients with tuberculosis.It is associated with increased antibiotic use and prolongs the duration of hospital stay.This is a symptom of a complex and serious clinical state.The etiology of high fever in tuberculosis is multifactorial,the most common one is with other bacterial infection or fungal infection.These findings suggest that improved protocols for fever management in those patients would contribute to an rational antibiotic use in tuberculosis. |
PDF Full Text Request