Spotted fever (SF), caused by Rickettsia conorii, and murine typhus (MT), caused by Rickettsia typhi, are endemic in Israel and are a common source of undifferentiated febrile illness.1,2
The clinical presentation of rickettsiosis in children is often confused with viral infection, suggesting that many cases are not recognized.3 Furthermore, differentiating SF from MT on clinical grounds is challenging, because exposure to tick (SF vector) or flea (MT vector) bites is rarely recalled or noticed, and both diseases share signs and symptoms of fever, headache, malaise, gastrointestinal symptoms and rash. Rash, in particular is highly variable, documented in over 90% of SF but ranges between ~20% and 70% in MT. Frequent laboratory abnormalities include leukopenia, thrombocytopenia, hyponatremia and elevated liver enzymes.2–5
The serologic diagnosis of rickettsiosis bears several limitations, including poor sensitivity, particularly when using a single test (rather than sequential testing for convalescent antibodies), and poor specificity, driven by cross-reactivity (CR) between rickettsial antigens.6,7 Additionally, background seroprevalence may confound unrelated illness or reinfection with rickettsiosis. These limitations lead physicians to start empiric treatment before the disease is confirmed.
Our aim was to assess the performance of serology tests in clinically suspected acute rickettsioses among children.
This was a retrospective, cohort study, conducted at the Soroka University Medical Center, and included all children ≤18 years old, who had serologic testing for rickettsioses, between 2015 and 2018. The study was approved by the Soroka University Medical Center Ethics Committee.
In southern Israel, 2 distinct ethnic populations live side by side: the Moslem Bedouin and the Jewish population. Higher rates of rickettsiosis were previously reported among the Bedouin population.2
Blood serology was evaluated by using immunofluorescent assay test kit for IgM and IgG for R. typhi and R. conorii, as described previously.8 The demonstration of IgM or IgG antibodies at a titer ≥1:100 was considered diagnostic.
Cases of concomitant MT and SF antibodies were considered CR. If IgM antibodies for both antigens were present at the same titer levels, a diagnosis was made based upon IgG titers; in inconclusive cases, a diagnosis was based on a second IgG test titers (convalescent).
The following data were obtained: demographic details, clinical manifestation, laboratory results, treatment and outcome (hospitalization, admission to the pediatric intensive care unit and mortality).
Descriptive statistical univariate analyses were calculated using SPSS 24 (IBM SPSS Statistics for Windows, Version 24.0. IBM Corp., Armonk, NY). Univariate analysis was performed using the two-tailed χ2 test, Fisher’s exact test or Student’s t test to assess differences between serology positive and negative cases of clinically suspected rickettsiosis. The threshold of statistical significance was P value <0.05.
Logistic regression models were used to evaluate demographic and clinical variables associated with serology-confirmed cases (the dependent variable being serology-positive rickettsiosis, including MT, SF and CR). These variables were selected from all tested variables statistically significant at the level of P < 0.2 in the univariate analyses. The P values were adjusted for multiple testing using the Benjamini–Hochberg method.
Overall, 963 episodes were identified: 33% (n = 322) had positive serology, and 67% (n = 641) had negative serology.
Positive Versus Negative Serology Episodes
Positive serology cases were associated with older age (71% vs. 43%, P = 0.001), Bedouin ethnicity (91% vs. 79%, P = 0.001) and seasonality [warm months (73% vs. 61%, P = 0.001)] (Table 1).
Table 1. -
Serologic Diagnosis of Clinically Suspected Acute Rickettsiosis in Children: Comparison of Positive and Negative Serology Cases
||Positive Serology (N = 322)
||Negative Serology (N = 641)
|Mean ± SD (range)
||11.1 ± 4.7 (0.4–18.4)
||7.7 ± 5.4 (0.1–18.2)
| ≥8, n (%)
| >1, n (%)
|Gender, n (%)
|Ethnicity, n (%)
||29 (9.0 )
|Seasonality*, n (%)
| Warm months
| Fever (≥38 °C), n (%)
| Mean temperature ± SD
||39 °C ± 1.1
||39.1 °C ± 1.2
| Presence of rash, n (%)
| CNS involvement, n (%)
| WBC (mm3), mean ± SD
||7.4 ± 4.3 (N = 321)
||9.8 ±6.3 (N = 638)
| WBC ≤ 5000/mm3, n (%)
| WBC ≥ 15,000/mm3, n (%)
| Platelets (mm3), mean ± SD
||208.5 ± 108 (N = 321)
||237.5 ±133.6 (N = 638)
| Platelets (mm3) ≤ 150, n/N (%)
| Sodium (mEq/L), mean ± SD
||134.4 ± 3.3 (N = 313)
||134.9 ±3.2 (N = 629)
| Sodium ≤ 135 mEq/L, n/N (%)
| GOT (units), mean ± SD
||70 ± 287.6 (N = 286)
||47.8 ±50.5 (N = 559)
| GOT ≥ 40 U, n/N (%)
| GPT (units); mean ± SD
||44.6 ± 158.1 (N = 296)
||34.6 ±66.6 (N = 583)
| GPT ≥ 40 U, n/N (%)
| CRP (mg/dL), mean ± SD
||6.6 ± 7.5 (N = 257)
||6.6 ± 7.5 (N = 531)
| CRP ≥ 7 mg/dL, n/N (%)
| Any treatment, n (%)†
| Doxycycline, n/N (%)
| Hospitalization, n (%)
| Hospitalization duration (d), mean ± SD
||4.2 ± 2.6 (N = 218)
||5.6 ± 6.4 (N = 447)
| Hospitalization duration ≥ 4 d, n/N (%)
| PICU, n (%)
| Mortality, n (%)
*Seasons: warm months—Summer and Fall (June to November).
CNS indicates central nervous system; GOT, glutamic-oxaloacetic transaminase; GPT, glutamic-pyruvic transaminase; PICU, pediatric intensive care unit; WBC, white blood cells.
Fever was documented in 90% and 86% of seropositive and seronegative cases, respectively. Rash was noted in similar rates in the positive and negative serology groups (29% vs. 31%, P = 0.43), with similar distribution of maculopapular manifestation (95% vs. 90%, P = 0.19), petechial manifestation (15% vs. 11%, P = 0.33) and palmar/plantar involvement (57% vs. 56%, P = 0.9). In contrast, petechial manifestation (15% vs. 11%, P = 0.33) and face involvement (19% vs. 9%, P = 0.055) were more common with seronegative cases.
Positive serology cases were associated with leukopenia, thrombocytopenia, hyponatremia, elevated glutamic-pyruvic transaminase and elevated glutamic-oxaloacetic transaminase. In contrast, leukocytosis was associated with negative serology.
Regarding treatment, positive serology episodes were associated with higher rates of antibiotic treatment, specifically doxycycline.
Hospitalization and mortality rates were similar when comparing both groups. In contrast, a prolonged hospitalization duration of ≥4 days and intensive care unit admission were associated with negative serology.
Age ≥8 years [odds ratio (OR) = 2.693; 95% CI: 1.968–3.686], seasonality (warm months) (OR = 1.517; 1.094–2.104) and leukocyte count of 5000–15,000/mm3 (OR = 2.247; 1.317–3.832) were associated with positive serology. In contrast, male gender (OR = 0.721; 0.529–0.983), Jewish ethnicity (OR = 0.428; 0.270–0.678), hospitalization duration ≥4 days (OR = 0.690; 0.504–0.945) and lack of doxycycline treatment (OR = 0.524; 0.373–0.737) were negatively associated with positive serology.
MT Versus SF Positive Serology
Of all positive serology cases (n = 322), 55%, 39% and 7% were positive for MT, CR and SF, respectively. Positive IgM titers were noted in 93% versus 67% in MT versus SF, respectively (P = 0.02). Positive IgG titers were noted in 28% versus 43% in MT versus SF, respectively (P = 0.15).
MT cases were associated with older age (11.5 ± 4.9 vs. 7.8 ± 5.8 years, P < 0.001). Gender, ethnicity and seasonality were similar in both MT and SF (Table, Supplemental Digital Content 1, https://links.lww.com/INF/E510).
There was a trend toward lower rates of rash in MT than in SF (25% vs. 43%, P = 0.08). Maculopapular manifestation was most frequent and appeared in similar rates in MT and SF (95% and 100%, respectively), while petechial rash (78% vs. 7%, P < 0.001), face involvement (33% vs. 7%, P = 0.053) and palmoplantar involvement (78% vs. 61%, P = 0.46) were more common in SF. Fever was documented in 86% and 90% of MT and SF, respectively.
Rates of leukocytosis, hyponatremia and elevated C-reactive protein (CRP) levels were higher in SF cases than those noted at MT. Rates of thrombocytopenia and elevated liver enzymes were similar in MT and SF.
Regarding treatment, similar rates of antibiotic treatment (47% vs. 57%) and specifically doxycycline (98% vs. 92%, P = 0.33) were reported in MT and SF, respectively.
Rates of hospitalization, intensive care unit admission and mortality were similar when comparing MT and SF. Hospitalization duration ≥4 days was associated with SF episodes (73% vs. 47%, P = 0.05).
Most CR episodes had higher antibody titers for MT than for SF (IgM—70% vs. 10%, IgG—50% vs. 3%). Of the 124 CR episodes, 81 (65%) were diagnosed as MT, 9 (7%) were diagnosed as SF and for the remaining 34 episodes (27%), no final diagnosis was determined.
In the current study, we assessed serology test performance in clinically suspected acute rickettsioses among children. Overall, a third of all serology tests were positive for rickettsioses. Most positive serology tests were associated with MT, while only a small minority (~10%) were associated with SF. This is even more striking when considering that an in-hospital setting usually has an overrepresentation of severe diseases, rather than mild or self-limiting ones.
The low rate of SF in the current study was evident in both overall and CR cases (which comprised 39% of all positive serology cases). Furthermore, the relatively low rate (29%) of rash among the positive serology cases group probably indicates that most cases were not SF in this group, because rash is reported in >90% of SF, while reported rates in MT range from 6.6% to 80%.6,9,10 The low rate of SF can be explained by either a true low rate of SF in the pediatric population of southern Israel or by an underdiagnosis of SF when assessed only by a single serology testing. Notably, similar seroprevalence rates for both diseases were previously reported in our region.4
It is our speculation that SF underdiagnosis derived from our study design, rather than from a true low rate of SF in our region. Because SF is a more severe disease, presenting as an acute febrile disease, it is possible that many SF cases are in fact often missed by early serology testing. This possibility is highlighted by the higher positivity rates of IgM among MT compared with SF cases.
Serologic testing of rickettsioses is routinely used in the clinical setting, yet our findings suggest that during acute disease, SF episodes may have been underdiagnosed or misdiagnosed, increasing the risk for delayed treatment and complications. The practice of single serology testing, without evidence of seroconversion, does not allow confirmation of diagnosis or the establishment of the true rate of MT to SF rickettsioses. Undefined background seroprevalence rates, CR, lack of data regarding successive serology tests and the possibility of dual infection (MT and SF), all underscore the limited benefit of serology testing in these settings. The introduction of new molecular methods, primarily of the polymerase chain reaction technique, may play a crucial role in current and future management of clinical rickettsioses.7
The main strengths of this study include the large number of episodes from a large tertiary medical center. The limitations of our study lie in its retrospective nature and the fact that the study was conducted in a single center.
In conclusion, Bedouin ethnicity, leukopenia, hyponatremia and elevated liver enzymes were associated with positive serology episodes. Among those, rash rates were low, MT predominated over SF and IgM positivity rate was higher in MT, compared with SF. These findings suggest that during acute rickettsiosis disease presentation in the hospital setting, serology testing may reliably identify MT infection but possibly fail to diagnose SF.
1. Raoult DBennet J, Dolin R, Blaser MChapter 186: Introduction to rickettsioses, ehrlichioses, and anaplasmoses.In: Mandell, Douglas, and Bennett’s Principles and Practice of Infectious Diseases20199th edElsevier. Inc24952498
2. Mumcuoglu KY, Ioffe-Uspensky I, Alkrinawi S, et al.Prevalence of vectors of the spotted fever group rickettsiae and murine typhus in a Bedouin town in Israel.J Med Entomol200138458461
3. Perez-Avraham G, Yagupsky P, Schlaeffer F, et al.Zoonotic infections as causes of hospitalization among febrile Bedouin patients in southern Israel
.Trans R Soc Trop Med Hyg200195301303
4. Shaked Y, Samra Y, Maeir MK, et al.Murine typhus and spotted fever in Israel in the eighties: retrospective analysis.Infection198816283287
5. Antón E, Font B, Muñoz T, et al.Clinical and laboratory characteristics of 144 patients with mediterranean spotted fever.Eur J Clin Microbiol Infect Dis200322126128
6. La Scola B, Raoult DLaboratory diagnosis of rickettsioses: current approaches to diagnosis of old and new rickettsial diseases.J Clin Microbiol19973527152727
7. Paris DH, Dumler JSState of the art of diagnosis of rickettsial diseases: the use of blood specimens for diagnosis of scrub typhus, spotted fever group rickettsiosis, and murine typhus.Curr Opin Infect Dis201629433439
8. Goldwasser RA, Steiman Y, Klingberg W, et al.The isolation of strains of rickettsiae of the spotted fever group in Israel and their differentiation from other members of the group by immunofluorescence methods.Scand J Infect Dis197465362
9. Shalev H, Raissa R, Evgenia Z, et al.Murine typhus is a common cause of febrile illness in Bedouin children in Israel.Scand J Infect Dis200638451455
10. Tsioutis C, Zafeiri M, Avramopoulos A, et al.Clinical and laboratory characteristics, epidemiology, and outcomes of murine typhus: a systematic review.Acta Trop20171661624