Aetiology of acute febrile illness: a multicentre study from the province of Kerala in southern India

Abstract

Acute febrile illness with varied aetiology but similar symptoms is common in tropical countries. This prospective, multicentre study was conducted in selected centres in the province of Kerala in India principally to analyse the aetiology of acute febrile illnesses in adult patients over the course of one year. Overall, 1324 patients were included in the study. The most common cause was dengue in 576 patients (43.5%). In 396 (29.9%), the exact aetiology could not be identified. Other causes, in order, were leptospirosis, enteric fever, malaria, respiratory tract infection, urinary tract infection and typhus. When such a wide variation with a significant number of ‘indeterminate’ cases exists, especially in such a small area and with limited resources, the onus is on public health authorities to draw up an ‘easy-to-use algorithm’ to tackle epidemics of febrile illness, particularly in the monsoon season.

Keywords

Asia diagnosis public health

Introduction

Acute undifferentiated febrile illnesses (AUFI) are common in the tropics. However, in a large country such as India there is variation even from area to area.^1–4 Kerala is one such province in the southern part of India.

Acute fever cases in which no focus of infection can be elucidated by history, physical examination or routine investigations in the early stages are referred to as AUFI. Common causes of AUFI described from the tropics include dengue, malaria, leptospirosis, scrub typhus, enteric fever, viral hepatitis, influenza, chikungunya and hanta virus infection. In the western world, acute febrile illness is mainly due to non-specific viral illnesses.

Many of the tropical diseases causing AUFI have similar presentations. Patients undergo various investigations which inevitably delay treatment and increase the cost. In spite of recurring epidemics, scientific data about these epidemics remain scanty. It is in this background that a study on AUFI was planned in multiple centres in Kerala.

Patients and methods

Ours was a prospective observational study conducted at Government Medical College Hospital, Thrissur and M.O.S.C Medical College Hospital, Kolenchery. Both these hospitals serve three adjoining districts (Figure 1). The period of study was 15 July 2014 to 14 July 2015.

Patients aged >13 years with an axillary temperature of 38℃ or higher of a duration of <2 weeks were included in the study. Excluded were critically ill patients and where accompanying persons were unable to provide details and trauma-related cases of fever. Patients having single organ involvement such as pneumonia, acute respiratory tract infection and urinary tract infection with symptoms and signs pertaining to the same were likewise excluded; however, those who did not have any specific symptoms were included.

To record clinical findings, investigators used a standardised case report form. The primary objective was to identify causes of acute fever using pre-set diagnostic criteria and laboratory tests. Data were entered into an Excel spreadsheet and analysed with Epi info version 7 statistical software. The protocol and related documents were approved by the regional ethics committee of both institutions.

Results

During the study period, 1324 inpatients with a diagnosis of AUFI were recorded (837 men [63.2%]; age range = 13–92 years). The majority of AUFI occurred among the 20–49-year age group (Table 1).

Table 1.

Age and gender of patients with acute febrile illness.

Age group (years)	Male (n = 837)	Female (n = 487)	Total (n = 1324)
<20	93 (11.11)	39 (08.01)	132 (09.96)
20–29	197 (23.53)	97 (19.92)	294 (22.21)
30–39	155 (18.51)	97 (19.92)	252 (19.03)
40–49	161 (19.23)	111 (22.79)	272 (20.54)
50–59	115 (13.73)	82 (16.83)	197 (14.88)
60–69	64 (07.65)	43 (08.83)	107 (08.08)
70–79	47 (05.62)	16 (03.29)	63 (04.76)
80–89	04 (00.48)	02 (00.41)	06 (00.45)
>90	01 (00.12)	00 (0)	01 (00.08)

Values are presented as n (%).

Headache was the most common symptom and was present in 549 patients (41.5%). Other common features included vomiting (485, 36.6%) followed by myalgia (299, 22.6%). A total of 576 (43.5%) patients had dengue fever and 238 (18.0%) were diagnosed with leptospirosis. In 396 patients (29.9%), the cause of fever was undiagnosed. Other causes of fever included malaria, enteric fever, respiratory tract infection, urinary tract infection and typhus (Table 2). All the aetiologies of fever were more common in men except for urinary tract infection. Thrombocytopenia was seen in dengue, leptospirosis, malaria, enteric fever and typhus. The mean platelet count was 76,500 /uL in dengue while in other causes of fever it was >100,000/uL. Aspartate transaminase (AST) elevation up to 9796 U/L was observed in dengue; AST was higher than alanine transaminase (ALT) in dengue.

Table 2.

Aetiology.

Diagnosis	n	%
Dengue	576	43.50
Undiagnosed fever	396	29.91
Leptospirosis	238	17.98
Enteric fever	61	04.61
Malaria	38	02.87
Urinary tract infection	7	00.53
Respiratory tract infection	6	00.45
Typhus	2	00.15
Total	1324	100.00

Twenty (1.5%) patients with fever died while in the hospital. Among these, 11 (55%) death were a result of leptospirosis (Table 3). Of deaths, 73% occurred in patients who were admitted during the monsoon and immediate post-monsoon season, whereas only 27% of the mortality occurred in the summer season. Acute kidney injury (odds ratio [OR] = 7.94; P = 0.03), hepatic dysfunction (OR = 11.03; P = 0.007) and acute respiratory distress syndrome (OR = 20.59; P = 0.0002) were predictors of mortality in acute febrile illness. Low platelet counts were also an independent predictor of mortality.

Table 3.

Outcome.

Diagnosis	Total patients	Deaths (n = 20) (% of total deaths)
Dengue	576	1 (5)
Leptospirosis	238	11 (55)
Undiagnosed fever	396	8 (40)

Discussion

The syndrome of acute febrile illnesses may have varied aetiology and may be influenced by local outbreaks which vary from continent to continent, nation to nation and even area to area in the same country. In this hospital-based prospective multicentre study, the first of its kind from Kerala, patients admitted with acute febrile illness were analysed. The major causes of acute febrile illness were dengue, ‘undiagnosed fever’, leptopirosis, enteric fever and malaria.

Dengue was the leading cause and has been documented in Kerala as early as 1979.⁵ However, in recent years, there has been a steep increase in its incidence. During the epidemic of 2013, 75,808 cases and 193 deaths were reported in India. Kerala ranked second in dengue mortality.⁶ In a significant group (n = 396, 29.9%), the cause of fever was undiagnosed. Owing to limited resources, an evaluation of the aetiology could not be done in these cases. This is a distinct limitation of our study. We hypothesise that most of these cases may be ‘viral fever’ or milder forms of dengue or leptospirosis where serological positivity (antibody formation) would be detected only later in the disease. In studies of AUFI from other tropical regions of the world, this phenomenon of ‘undiagnosed fever’ is commonly reported.^1,7–9 Leptospirosis, the third leading cause of acute febrile illness in our study, has been reported from Kerala since as early as 1990. Rapid urbanisation and construction activities have led to an ever-increasing number of cases since then.¹⁰ There is a higher incidence during the monsoon and immediate post-monsoon period.^2,10 Malaria is also emerging as an important cause of fever in Kerala, owing to a migrant population from north India coming as construction and agricultural field workers.¹¹

Even for a small area like Kerala, the variation in the aetiological profile that spans time and region has been highlighted in this study, when comparison is made to an earlier single-centre study from one of the participating centres of this study.² In that study, ‘presumed viral fever’ formed the single largest group (51.5%) followed by leptospirosis, dengue, typhoid and malaria.

Since symptomatology in acute febrile illness may be varied, they have a poor predictive value in determining aetiology.⁹ This study was no exception. Likewise, leucopenia and thrombocytopenia may occur in several common diseases and transaminase levels are none too specific either.

AUFI continues to be a challenge to tropical countries. An ‘easy-to-follow’ algorithm may mitigate the challenge to diagnosis. Such an algorithm has been suggested in an earlier study² from one of the participating centres of this study. However, such algorithms are limited by variation in aetiology, resources and commitment of the health provider. Hence, the algorithm needs to be modified / tailor-made to suit the needs of a particular centre and validated periodically.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by a grant from the Indian Council of Medical Research (ICMR): IRIS No. 2012-27520/F. No. 5/6/3/12-RM.

References

Mittal

Ahmad

Agarwal

et al.

Aetiologies of acute undifferentiated febrile illness in adult patients - an experience from a tertiary care hospital in northern India. J Clin Diagn Res 2015; 9: DC22–24.

Ittyachen

Ramachandran

. Study of acute febrile illness: a 10-year descriptive study and a proposed algorithm from a tertiary care referral hospital in rural Kerala in Southern India. Trop Doctor 2015; 45: 114–117.

Chrispal

Boorugu

Gopinath

et al.

Acute undifferentiated febrile illness in adult hospitalized patients: the disease spectrum and diagnostic predictors - an experience from a tertiary care hospital in South India. Trop Doctor 2010; 40: 230–234.

Joshi

Colford

Jr Reingold

et al.

Nonmalarial acute undifferentiated fever in a rural hospital in central India: diagnostic uncertainty and overtreatment with antimalarial agents. Am J Trop Med Hyg 2008; 78: 393–399.

Sreenivasan

Rodriques

Venkateshan

et al.

Isolation of dengue virus from Trichur district (Kerala State). Indian J Med Res 1979; 69: 538–541.

Government of India and Ministry of Health and Family Welfare. National Vector Borne Disease Control Program. Annual report 2014–15. New Delhi: Ministry of Health and Family Welfare, 2014. Available at: http://www.nvbdcp.gov.in/Doc/Annual-report-NVBDCP-2014-15.pdf.

Manock

Jacobsen

de Bravo

et al.

Etiology of acute undifferentiated febrile illness in the Amazon basin of Ecuador. Am J Trop Med Hyg 2009; 81: 146–151.

Leelarasamee

Chupaprawan

Chenchittikul

et al.

Etiologies of acute undifferentiated febrile illness in Thailand. J Med Assoc Thai 2004; 87: 464–472.

Abhilash

Jeevan

Mitra

et al.

Acute undifferentiated febrile illness in patients presenting to a tertiary care hospital in south India: clinical spectrum and outcome. J Glob Infect Dis 2016; 8: 147–154.

10.

Pappachan

Sheela

Aravindan

. Relation of rainfall pattern and epidemic leptospirosis in the Indian state of Kerala. J Epidemiol Community Health 2004; 58: 1054–1054.

11.

Antony

Celine

Chacko

. Staging back of Malaria in Kerala, India: A Retrospective study. Int Res J Social Sci 2013; 2: 42–46.