Differentiating Spotted Fever From Scrub Typhus Based on Clinical and Laboratory Features: A Retrospective Cohort Study From South India

Background
Scrub typhus and spotted fever are the two most common rickettsial infections in India. Both conditions differ, however, in terms of epidemiological risk, clinical presentation, and prognosis. We aimed to compare the clinical and laboratory characteristics of patients diagnosed with scrub typhus or spotted fever to identify disease indicators that are useful in low-resource endemic settings.

Methods
We conducted a retrospective cohort study among adults admitted with undifferentiated fever lasting 5–15 days at a tertiary care hospital in coastal Karnataka, India, between January 2017 and December 2023. Probable spotted fever and probable scrub typhus were diagnosed by single positive serology, using the Weil–Felix test and IgM ELISA, respectively. Patients with coinfection between the two study conditions or with any other confirmed pathogen were excluded from the analysis. Binary logistic regression was performed with spotted fever as the dependent variable and scrub typhus as the referent category.

Results
Of 1045 patients with undifferentiated fever, 94 met the study criteria for probable spotted fever and 643 for probable scrub typhus. Probable spotted fever was observed throughout the year, whereas probable scrub typhus showed a marked monsoon and post-monsoon predominance, from June to November. Rash (39.4% vs 11.0%, P < .001) and leucocytosis (46.8% vs 26.6%, P < .001) were significantly more frequent in probable spotted fever than in probable scrub typhus. In contrast, splenomegaly (21.3% vs 37.5%, P = .002), lymphadenopathy (2.1% vs 11.8%, P = .004), lung involvement (34.0% vs 62.2%, P < .001), and hypoalbuminemia (55.6% vs 75.7%, P < .001) were less frequent in probable spotted fever than in probable scrub typhus. In multivariable analysis with spotted fever as the outcome, rash (adjusted odds ratio [aOR], 5.69, P < .001) and leucocytosis (aOR, 3.67, P < .001) were independently associated with higher odds of probable spotted fever. In contrast, monsoon or post-monsoon presentation (aOR, 0.48, P = .006), splenomegaly (aOR, 0.45, P = .008), lymphadenopathy (aOR, 0.09, P = .002), lung involvement (aOR, 0.29, P < .001), and hypoalbuminemia (aOR, 0.45, P = .003) were associated with lower odds of probable spotted fever relative to probable scrub typhus. There was no significant difference in day-7 mortality between the two groups.

Conclusions
Probable scrub typhus and probable spotted fever can be differentiated by a set of epidemiological, clinical, and first-line laboratory variables. These findings should improve the bedside diagnostic approach while awaiting further multicenter prospective validation.