Morbidity and Functional Outcomes Following Rocky Mountain Spotted Fever Hospitalization—Arizona, 2002–2017

Abstract

Background

Rocky Mountain spotted fever (RMSF) is a deadly tickborne disease disproportionately affecting Arizona tribal communities. While the acute clinical effects of RMSF are well-documented, more complete understanding of the long-term health consequences is needed to provide guidance for providers and patients in highly impacted areas.

Methods

We performed a retrospective review of hospitalized RMSF cases from 2 tribal communities in Arizona during 2002–2017. Medical records from acute illness were abstracted for information on clinical presentation, treatment, and status at discharge. Surviving patients were interviewed about disease recovery, and patients reporting incomplete recovery were eligible for a neurologic examination.

Results

Eighty hospitalized cases of RMSF met our inclusion criteria and were reviewed. Of these, 17 (21%) resulted in a fatal outcome. Among surviving cases who were interviewed, most (62%) reported full recovery, 15 (38%) reported ongoing symptoms or reduced function following RMSF illness, and 9 (23%) had evidence of neurologic sequelae at the time of examination. Sequelae included impaired cognition, weakness, decreased deep tendon reflexes, seizures, and cranial nerve dysfunction. Longer hospitalization (25.5 days vs 6.2 days, P < .001), a higher degree of disability at discharge (median modified Rankin score 1 vs 0, P = .03), and delayed doxycycline administration (6.2 days vs 4.1 days, P = .12) were associated with long-term sequelae by logistic regression.

Conclusions

Although the etiology of sequelae is not able to be determined using this study design, life-altering sequelae were common among patients surviving severe RMSF illness. Delayed administration of the antibiotic doxycycline after day 5 was the strongest predictor of morbidity.

Rocky Mountain spotted fever (RMSF) is a deadly tickborne disease. We identified that 38% of surviving, hospitalized individuals with RMSF reported ongoing symptoms after recovery and 23% had evidence of neurologic sequelae. The presence of sequelae was associated with delayed treatment.

Rocky Mountain spotted fever (RMSF) is a potentially deadly tickborne disease caused by the bacterium Rickettsia rickettsii. RMSF is a rapidly progressing illness; bacteria invade endothelial cells throughout the body resulting in disrupted organ perfusion, sepsis, and sometimes death [1]. Prior to antibiotic therapy, case fatality rates from RMSF were as high as 60% [2, 3]. However, with advances in critical care and widespread availability of tetracycline class antibiotics, case fatality rates have dropped significantly to 5%–10% in the United States [4–6]. Delayed treatment beyond day 5 of illness is the single most important predictor of severe and fatal outcomes from RMSF [6–8].

RMSF is characterized initially as an acute febrile illness with nonspecific symptoms such as fever, headache, and muscle pain. A disseminated rash usually appears between day 2 and day 4 of illness. Beyond day 5 of illness, patients may begin experiencing severe abdominal pain, respiratory distress, and acute neurologic signs and symptoms. Neurologic symptoms are thought to result from blood vessel injury in the central nervous system or by the resulting inflammatory process [1, 9]. Early accounts of neurologic impairment included persistent changes in neurologic function, loss of motor control, and behavior change [10, 11]. Subsequent case series similarly documented neurologic dysfunction years after acute disease recovery, although many included small sample sizes and varying types of neurologic evaluations [9, 12, 13].

Arizona tribal communities have experienced epidemic levels of RMSF since 2002, with some areas reporting incidence rates 150 times higher than the national average [14]. In the years following the initial surge in cases, families and providers began reporting persistent impairment among patients surviving acute illness. In this study, we reviewed hospitalized cases in a large cohort to better understand sequelae following RMSF, the frequency and duration with which symptoms persist, potential for improvement, and the expected degree of disability.

METHODS

We used a convenience sample of individuals hospitalized with RMSF during 2002–2017 from 2 highly impacted communities in Arizona. Cases were identified using data reported to the Arizona Department of Health Services. For the purposes of this study, long-term sequelae (LTS) are defined as symptoms reported by patients or signs identified on exam following acute RMSF illness and lasting >1 year. Data for this study were obtained from 3 separate sources: (1) abstraction of medical record from acute RMSF illness; (2) interview of surviving patients, and (3) neurologic exam of patients reporting persistent sequelae. All cases meeting the case criteria were included in the medical record abstraction, and all individuals who could be contacted were eligible for patient interview.

Inclusion Criteria

Cases were included based on criteria adapted from the Council of State and Territorial Epidemiologists case definition for spotted fever rickettsiosis (including RMSF) used in national disease surveillance [15]. Clinical criteria include fever with 1 or more of the following: rash, eschar, headache, myalgia, anemia, thrombocytopenia, or elevated hepatic transaminases. Afebrile cases with laboratory evidence and a clinical presentation otherwise consistent with RMSF were also included in this analysis. Only hospitalized individuals were included. Additionally, cases were required to have at least 1 of the following pieces of laboratory evidence: (1) elevated immunoglobulin G (IgG)–specific antibodies reactive to R rickettsii by indirect immunofluorescence antibody (IFA) assay with titer value ≥1:64; (2) detection of R rickettsii by polymerase chain reaction (PCR) assay; or (3) demonstration of spotted fever group Rickettsia (SFGR) antigen by immunohistochemistry. Serologic cross-reactivity among SFGR is known to occur, but all cases of spotted fever rickettsiosis from these 2 tribal communities with PCR confirmation have been speciated to R rickettsii. Therefore, RMSF is presumed to be the predominant cause of spotted fever rickettsiosis in this area.

Medical Record Abstraction

Clinical information was abstracted from the inpatient record during acute illness. Data collected included signs and symptoms of acute illness, neuroimaging studies, assessments on the severity of illness, time to treatment, comorbid conditions at the time of admission (including previously diagnosed neurologic dysfunction), and condition at the time of discharge (including disposition, documented neurologic status, referrals for follow-up or rehabilitation, and any identified disability). Sign and symptom information were based on reviews of the literature. Signs were defined as objective clinical or laboratory evidence; symptoms were defined as subjective physical or mental experiences (such as headache or anxiety) possibly resulting from disease process as reported in the medical record via patient history. Fever was defined as a temperature ≥38°C (100.4°F) or reported fever by the patient or caretaker. Severe illness was defined as 1 or more of the following: evidence of shock or use of vasopressors, acute respiratory distress syndrome, multiorgan failure, renal failure, cerebral edema, coma, digital necrosis, severe thrombocytopenia (<50 × 10³ platelets/µL), or use of mechanical ventilation. The modified Rankin scale was estimated by trained abstractors based on information provided in the medical record to provide an objective measure of neurologic disability at the time of discharge [16].

Patient Interviews

Surviving RMSF cases who could be contacted were eligible for interviews. Surveys included perceived disease recovery and functional capacity (using a 5-point Likert scale) before and after RMSF illness. Surveys were administered by a trained member of the tribal health department or hospital staff. Patients were asked about other neurologic diagnoses since their RMSF illness to control for unrelated, intervening causes of neurologic dysfunction. Individuals reporting decline in function from baseline, persistent symptoms, or difficulty completing tasks were eligible for a neurologic exam.

Neurologic Exam

Neurologic exams were conducted either by a board-certified neurologist or a licensed physician under the supervision of a board-certified neurologist. A standard neurologic exam included evaluation of mental status, cranial nerve function, motor and sensory function, reflexes, coordination, and gait. Additionally, cognitive impairment was further evaluated using the Montreal Cognitive Assessment (MoCA) [17], and results were compared to reference standards in the overall population [18]. Cognitive impairment was defined as scoring <26 on the MoCA. In cases where MoCA could not be completed, overall cognitive function was assessed during the mental status portion of the neurologic exam. Level of disability was again assessed using the modified Rankin scale at the time of neurologic assessment and compared to the score at the time of discharge. No imaging or laboratory tests were performed as part of this study. Patients were considered to have neurologic sequelae if the neurologic exam showed abnormal neurologic findings or if patients self-reported changes in function or the existence of neurologic abnormalities (such as seizures, changes in vision, hearing, bowel or bladder incontinence, or behavioral disturbances) at the time of neurologic exam.

Ethical Considerations

This protocol was reviewed and approved by the Institutional Review Board (IRB) at the Centers for Disease Control and Prevention (CDC), the Phoenix Area Indian Health Service IRB, and the tribal council or appropriate tribal governing body at each of the participating locations. The University of Georgia and Arizona Department of Health Services IRBs deferred to the CDC's determination. This study was further evaluated for burden under the paperwork reduction act. The data collection tools were approved by the Office of Management and Budget (0920-1267).

Patient Consent Statement

As per approved IRB determination, patients were asked to provide informed consent for participation in the questionnaire and neurological exam. Patients were further asked to provide consent for their medical record data to be included in analysis at the time of interview.

Data Collection and Analysis

Data for each activity were recorded on paper forms and deidentified data were subsequently entered into an encrypted Microsoft Access database on a CDC secure server. Data collections from each of the 3 sources were reviewed for completeness by the site supervisor at the end of each day. Statistical analyses were performed using SAS version 9.4 (Cary, North Carolina). Demographic characteristics and key clinical indicators of the study population were shown using descriptive statistics. Bivariate associations were further assessed with the primary outcome of the presence or absence of LTS among surviving patients using Fisher exact tests, the Student t test for normally distributed data, and Mann-Whitney test for ordinal, nonnormally distributed data. Authors made a post hoc determination that doxycycline administration was considered to be delayed if therapy was started on or after day 5 of illness based on previous publications [4, 6, 8]. Unadjusted odds ratios (ORs) obtained through logistic regression are reported for key associations. Significance was assessed at P < .05.

RESULTS

Medical records from 80 hospitalized cases of RMSF with illness during 2002–2017 were evaluated based on those meeting inclusion criteria. Half of the patients (n = 40 [50%]) were female, and median age was 15 years (IQR, 4–41 years). Seventeen cases (21%) resulted in a fatal outcome.

Medical Record Review

Hospitalized patients in this cohort experienced a wide range of clinical signs and symptoms, with the majority reporting fever (59%), and rash (66%) at some point during their illness; history of tick exposure was less common (31%) (Table 1). Sixty-one percent of cases reported illness onset during July–October. Patients were hospitalized for a median of 4 days (IQR, 3–7 days; range, 1–60 days) with 40% requiring admission to an intensive care unit; severe illness was reported in 44% of cases. Twenty-eight (35%) individuals were discharged from the hospital with documented neurologic and psychiatric sequelae including weakness, pain, hearing loss, seizures, ataxia or balance problems, visual impairment, speech or swallowing dysfunction, coma, anxiety, or depression. Other, less commonly reported sequelae included paresthesia, incontinence, tremors, or hyporeflexia.

Table 1.

Characteristics of Hospitalized Cases of Rocky Mountain Spotted Fever—Arizona, 2002–2017 (N = 80)

Variable	No. (%)
Median age, y (IQR)	15 (5–41)
Female sex	40 (50)
Month of illness onset
January–March	13 (16)
April–June	14 (17)
July–September	37 (46)
October–December	16 (20)
Received doxycycline <5 d from illness onset	35 (53)
Median days to first doxycycline administration (IQR)	4 (2–6)
Comorbid medical condition
Diabetes mellitus	15 (19)
Neurologic condition	13 (16)
Psychiatric illness	12 (15)
Report of tick bite	25 (31)
Rash	53 (66)
Fever	47 (59)
Admission to ICU	32 (40)
1 or more severe signsa	35 (44)
Neurologic or psychiatric sequela at discharge	28 (35)
Disposition at discharge
In-hospital death	17 (21)
Nonfatal impairmentb	54 (68)
No impairment	9 (11)

Abbreviations: ICU, intensive care unit; IQR, interquartile range.

^aOne or more of the following: evidence of shock or use of vasopressors, acute respiratory distress syndrome, multiorgan failure, renal failure, cerebral edema, coma, digital necrosis, severe thrombocytopenia (<50 × 10³ platelets/µL), or use of mechanical ventilation.

^bBased on modified Rankin score 1–5.

Patient Interview

Twelve individuals consented for interview but did not approve of inclusion of their medical record information; their data are only included in the interview section. Other patients from the medical record review were deceased at the time of interview (n = 22), lost to follow-up (n = 27), or refused to participate (n = 3). Therefore, 40 survivors of acute infection (63%) were successfully contacted following their illness for the patient questionnaire. Most patients interviewed (62%) experienced full recovery from their disease; 15 cases (38%) reported ongoing symptoms or reduced function following their RMSF illness. The median score of self-reported functionality on the 5-point Likert scale following RMSF illness was significantly lower among persons reporting incomplete recovery (4 [IQR, 2-4]) than among those who recovered completely (5 [IQR, 5-5]) (U test P = .01). A score of 5 indicated the individual was “perfectly able to function.” Among those who reported recovering from illness, the mean time to normal function was 12.7 days (range, 1–30 days).

Neurologic Examination

Fifteen individuals qualified for the neurologic exam based on interview responses. Twenty-three percent (9/40) of RMSF survivors reported symptoms or showed signs of LTS. Sequelae observed during the exam included impaired cognition; decreased deep tendon reflexes; cranial nerve dysfunction including visual impairment; weakness; abnormal gait, station, and coordination; and paresthesia (Tables 2 and 3). Additional LTS reported by patients at the time of neurologic exam included seizures, pain, photosensitivity, behavioral concerns, and fecal urgency. Six of the 9 cases with LTS had confirmatory laboratory evidence, with positive PCR, positive immunohisotchemistry (IHC), or 4-fold change in IgG by IFA. The 3 remaining cases had elevated IgG titers and clinical illness consistent with a recent rickettsial disease, but no convalescent titers were obtained. Median age of patients with LTS was 38 years (IQR, 4–40 years) at the time of illness, and clinical evaluation was a median of 6 years (IQR, 1–11 years) following acute illness.

Table 2.

Description of Hospitalized Cases of Rocky Mountain Spotted Fever With Long-term Sequelae—Arizona, 2002–2017 (n = 9)

ID	Age at Acute Illness, y	Time From Illness Onset to Initiation of Doxycycline, d	Duration of Inpatient Stay, d	Significant Clinical Findings During Hospitalizationa	Results of Neurologic Diagnostic Testing During Hospitalization	Sequelae at Time of Discharge	Time from Discharge to Neurologic Exam, y	Long-term Sequelae
1	1	10	15	ARDS, altered mental status/encephalopathy, mechanical ventilation, DIC, petechial rash, cerebral infarcts	MRI: Punctate infarcts in inferior cerebellum and focal lesions in corpus callosum	Disorientation/altered mental status, hearing loss, visual impairment, seizures, dysarthria, dysphagia, weakness, tremors, ataxia, reported decline in function	6	Altered mental status (E); dysphagia (E); decreased muscle tone in extremities (E); MoCA unable to be assessed, but grossly impaired in all domains (E); seizures (S); behavioral concerns (S)
2	54	8	12	Diffuse macular rash	None performed	Weakness and numbness	1	Impaired cognitive function with challenges in abstraction, recall, and language (M); slow motor finger taps and foot taps (E); decreased reflexes (E); positive Babinski (E); fatigue (S)
3	57	2	4	SIRS	None performed	None	5	Impaired cognitive function with challenges in language, attention, and abstraction (M); decreased visual acuity (E); decreased reflexes in upper extremities and absent in lower extremities (E)
4	1	5	7	Organ failure, coma, altered mental status, shock, severe thrombocytopenia, papular rash, mechanical ventilation, global encephalopathy	CT: Normal	…	…	…
EEG: global encephalopathy (diffuse slowing and disorganization)	Seizures	10	Language impairment (M); abnormal visual acuity (E); decreased upper extremity reflexes (E); pain (S); delayed developmental milestones (S)	…	…	…	…	…
5	4	6	1	Diffuse maculopapular rash	None performed	None	1	MoCA unable to be assessed, attention issues (E, S); difficulties with problem solving compared to level expected for age (E); nonaphasic, nondysphasic speech issues (E, S)b
6	39	5	7	Petechial rash	None performed	None	11	Paresthesia and numbness in lower and upper extremities (E); decreased reflexes in upper extremities and absent in lower extremities (E); fatigue (S); imbalance (S); fecal urgency (S)
7	40	5	4	Maculopapular rash, frontal cerebral infarcts	MRI: Punctate infarcts in peduncular and frontal white matter	None	1	Decreased reflexes (E); pain (S); photosensitivity (S)
8	4	5	60	Sepsis, ARDS, renal insufficiency, multiorgan failure, DIC, petechial rash, mechanical ventilation, cerebral atrophy	Multiple MRIs: Progressive cerebral atrophy.	Dysarthria, dysphagia, seizures, ataxia	11	Significantly impaired cognition (limited attention, language, recall and executive functions) (M); decreased reflexes bilaterally (E); bilateral weakness (E); expressive and receptive aphasia (E); decreased visual acuity (E); slow finger taps and foot taps (E); seizures (S)
9	39	11	7	Renal insufficiency, altered mental status, petechial rash	CT: Normal	Visual and auditory changes not specified	11	Abnormal language and recall (M); altered mental status (E); decreased visual acuity (E); decreased auditory function (cranial nerve VIII impairment) (E); unilateral numbness (E); absent and decreased reflexes (E); abnormal station (E); positive Romberg (E)

Abbreviations: ARDS, acute respiratory distress syndrome; CT, computed tomography; DIC, disseminated intravascular coagulation; E, clinical finding based on neurologic examination; EEG, electroencephalogram; M, abnormal Montreal Cognitive Assessment result; MoCA, Montreal Cognitive Assessment; MRI, magnetic resonance imaging; S, self-reported finding; SIRS, systemic inflammatory response syndrome.

^aExcluding comorbidities prior to Rocky Mountain spotted fever (RMSF).

^bUnclear speech, but not aphasic, some delays prior to RMSF, but family noted decline following RMSF.

Table 3.

Characterization of Neurologic Findings Observed During Examination Among Individuals With Long-term Sequelae—Arizona, 2002–2017 (n = 9)

Characteristic	Frequency Abnormala	Notes on Abnormal Findings
Mental status (MoCA)b	6/7 (86%)	Mean score = 19.1 (SD, 6.7)
Modified Rankin scale at examc	7/9 (78%)	Median score = 1 (IQR 1-1, 1; range, 1–3)
Reflexes	8/9 (89%)	Most had decreased deep tendon reflexes, 1 patient with hyperreflexia
Cranial nerve function	4/7 (57%)	CN II was most frequently affected showing decreased visual acuity. Facial sensory deficits (CN V) and hearing loss (CN VIII) were also noted. One additional individual could not be formally assessed for all cranial nerve function but had clear dysarthria.
Strength	4/7 (57%)	Reduced strength in the 3–4/5 range; all results were bilateral
Gait and station	3/9 (33%)	Wide-based gait or inability to walk on toes were the most common findings
Motor function	3/9 (33%)	Two patients experienced bilateral slowing of finger taps, 1 showed decreased tone in upper and lower extremities
Coordination	2/8 (25%)	Bilateral dysmetria on upper extremity past-pointing
Sensory function	2/9 (22%)	Numbness and paresthesia were reported
Language	1/9 (11%)	Receptive and expressive aphasia

Abbreviations: CN, cranial nerve; IQR, interquartile range; MoCA, Montreal Cognitive Assessment; SD, standard deviation.

^aFormal assessment of function could not be performed in some cases due to age and cooperation of patient. Unless otherwise noted, those not formally assessed were deemed to be grossly normal by the assessing physician.

^bNormative value for MoCA is 26.

^cNormative value for modified Rankin is 0.

Bivariate analyses of severe clinical manifestations during acute disease and subsequent development of LTS produced no statistically significant associations. Three of the 9 patients (33%) with LTS did not have neurologic sequelae recorded at the time of discharge. Patients with LTS had twice the odds of experiencing severe symptoms during acute illness, although this finding was not statistically significant (odds ratio [OR], 2.1 [95% confidence interval {CI}, .5–8.8]). Seizures were the most commonly reported sequela (33%) at discharge among individuals with LTS. Of the 3 individuals with LTS reporting seizures at discharge, 2 continued to experience seizures at 6 and 11 years postinfection. There were statistically significant differences in modified Rankin scores at discharge among patients with LTS and those without (median, 1 and 0, respectively; U test P = .07) (Table 4). Furthermore, patients with LTS experienced significantly longer duration of hospitalization (mean, 25.5 days vs 6.3 days; t test P < .001), and initiated doxycycline later than patients without LTS (6.2 days vs 4.2 days; t test P = .14) (Figure 1). In fact, patients with LTS had 19 times higher odds of initiating doxycycline after day 5 of illness (OR, 19.5 [95% CI, 2.3–167.9]).

Table 4.

Clinical Findings During Hospitalization for Rocky Mountain Spotted Fever Cases With and Without Long-term Sequelae Among Survivors—Arizona, 2002–2017 (n = 63)

Finding	Long-term Sequelae	No Long-term Sequelae	P Value
	(n = 9)	(n = 54)
Nonneurologic
Shock or vasopressor use	1 (11)	6 (11)	1
ARDS	2 (22)	5 (9)	.26
Multiorgan failure	2 (22)	6 (11)	.32
Renal insufficiency	2 (22)	8 (15)	.63
Necrosis	0 (0)	2 (4)	1
Severe thrombocytopenia	2 (22)	9 (17)	.65
Rash	8 (89)	33 (61)	.14
Supportive care procedures
Transfusion	2 (22)	4 (7)	.2
Hemodialysis	0 (0)	2 (4)	1
Amputation	0 (0)	0 (0)
Mechanical ventilation	3 (33)	6 (11)	.11
Neurologic
Coma	1 (11)	2 (4)	.38
Mental status changea	3 (33)	12 (22)	.43
Cerebral edema	0 (0)	1 (2)	1
Any neurological sequelae at discharge	6 (67)	21 (39)	.15
Any severe signsb	4 (44)	15 (28)	.43
Modified Rankin score, median (IQR)	1 (1)	0 (0–1)	.07
First doxycycline administration ≥5 d	8 (89)	16 (30)	.001c

Data are presented as No. (%) unless otherwise indicated.

Abbreviations: ARDS, acute respiratory distress syndrome; IQR, interquartile range.

^aAs reported from family or documented by medical staff at admission or during hospitalization.

^bOne or more of the following: evidence of shock or use of vasopressors, ARDS, multiorgan failure, renal failure, cerebral edema, coma, digital necrosis, severe thrombocytopenia (<50 × 10³ platelets/µL), or use of mechanical ventilation.

^cStatistically significant finding.

Figure 1.

Survival curve for death and disability among Rocky Mountain spotted fever cases—Arizona, 2002–2017. Includes only cases where medical records were available. *Includes those who never received doxycycline.

DISCUSSION

Acute Disease Findings

Our findings provide further evidence of the severe and fatal nature of RMSF. The case fatality rate was 21% in our cohort, which is higher than the estimated 5%–10% for the United States [4]. However, since we included only hospitalized cases, it is likely we captured those with more severe disease. High proportions of fatal outcome are consistent with reports from Mexicali and Sonora [19, 20]. The clinical descriptions of RMSF cases were similar to what has been reported in previous summaries from Arizona [6, 21]. Younger populations were heavily affected in this cohort (nearly 44% of cases were among children <10 years of age). Rash was the most frequent clinical sign (67%), similar to previous studies (68% [21]), though in a lower proportion than is reported nationally (88%–97%) [5, 22, 23]. Fever was reported in <60% of cases, which is the lowest proportion in the published literature and lower than was reported in the previous Arizona study (81%) [21]. Advanced illness with hypothermia at the time of first presentation may account for absence of fever in some cases as well as numerous accounts of antipyretic use at first visit. History of tick bite was infrequently reported (31%) in this cohort, lower than national reports (49%–60%) and the previous Arizona study (55%) [5, 21, 22]. Compared with the previous literature (15%), this study reports a profoundly higher proportion (35%) of individuals with neurologic sequelae at the time of discharge [5].

Recovery and Long-term Sequelae

Nearly half of the surviving individuals agreed to participate in the patient survey, providing key insight into disease recovery. Most patients reported full recovery with a mean time of 12 days to normal function. While time to defervescence and discharge have been previously reported, to our knowledge this is the first report on time to disease recovery. A substantial proportion (23%) of individuals had evidence of LTS from RMSF based on neurologic exam and reported neurologic symptoms. These estimates are consistent with previous studies on neurologic sequelae following RMSF (5%–56%), and represent the largest and most complete evaluation of LTS [9, 10, 12, 13]. Although a wide range of ages experienced LTS, the median age of at first illness among individuals with LTS (38 years) is significantly older than the median age of the overall hospitalized cohort (15 years). Cognitive impairments occurred frequently (86%), with the most impairment reported in areas of recall, language, and attention. Other common LTS included decreased deep tendon reflexes, decreased motor function, visual impairments and other cranial nerve dysfunction, weakness, changes to gait and station, and report of seizure activity. Paresthesias, language impairments, and abnormal coordination were reported, but less frequent.

Predictors of Long-term Sequelae

Individual clinical findings during hospitalization, presence of 1 or more elements of severe disease, and presence of neurologic sequelae at discharge were not significantly predictive of the development of LTS. Our ability to identify statistically significant predictors of LTS may, however, be limited by the small sample size of individuals with LTS. Delay in doxycycline administration is the strongest predictor of LTS by logistic regression. Survival analysis demonstrates precipitous changes in rates of adverse outcome (death or disability) if treatment is delayed. Furthermore, patients receiving doxycycline after day 5 of illness were significantly more likely to experience LTS and death. Increased morbidity and mortality following a delay in treatment past 5 days defines a critical time period, consistent with previous literature [6, 7]. Furthermore, the curve showing frequency of individuals with disability showed a steep decline once treatment was delayed past day 4 of illness onset, whereas mortality lagged by 1–2 days. This suggests a critical window where treatment delay may not cause death but can contribute substantially to disability and decreased functional outcome. Patients with LTS also experienced longer hospitalizations and were discharged with significantly more disability as denoted by the modified Rankin exam.

Many of the cases with documented LTS did not have reported neurologic sequelae at the time of discharge. This disconnect with disposition at the time of discharge with ultimate development of neurologic sequelae was previously documented [9]. Researchers theorized that neurologic damage was possible even in patients not experiencing severe illness or that sequelae were subclinical at the time of discharge, so development of neurologic sequelae following RMSF should not be ruled out solely based on discharge status. Neurologic imaging can be a helpful adjunct in the assessment of patients with RMSF; however, a thorough exam at discharge is equally valuable in identifying deficits and disability that may benefit from rehabilitation. If sequelae are identified at the time of discharge, supportive and rehabilitative care can be initiated to improve function.

Limitations

This study relies on convenience sampling of RMSF patients in specific Arizona tribal communities. While the findings are not yet proven to have external validity to other areas or to nonhospitalized cases, this cohort is the largest group of patients with RMSF known in the United States and provides substantial evidence of severe disability following RMSF. Similar studies are encouraged in other areas with RMSF to assess the external validity of these findings. We feel confident that those cases determined to have LTS represent true RMSF cases based on the high proportion of confirmatory laboratory evidence (4-fold change in IgG, PCR or IHC positive) and high single IgG tiers among those with supportive laboratory evidence. Comparisons were not made between cases with confirmatory laboratory evidence and those with only supportive laboratory criteria for this study; however, we assume a high proportion of confirmed cases based on previous studies of RMSF in Arizona compared to national surveillance data [21]. Despite clear inclusion criteria with clinical and laboratory components, it is possible that cases may be misclassified. Although only a few persons refused to participate, it is possible that persons experiencing disability may have been more likely to participate in interviews than those who were fully recovered, further inflating our estimates of disability. Medical record abstractions represent secondary reviews of information collected at the time of illness; the quality and completeness of records was highly variable. Neurologic exams were also subject to several biases. Although largely objective, neurologic exams may be subject to some interrater differences; we attempted to reduce that variability by having only 3 physicians conduct exams, and all were trained and supervised by a single board-certified neurologist. MoCA assessments are validated for English-speaking populations in individuals 8 years and older. All participants were offered the use of a certified health translator into their Native dialect. No individuals used this option; however, MoCA results may have been impacted by the administration of the test in their nonprimary language [24]. Finally, while an effort was made to identify intervening neurologic diagnoses during the patient survey, we are unable to rule out the contributions of other, more common causes of neurologic impairment, such as uncontrolled diabetes mellitus.

CONCLUSIONS

RMSF is a deadly and debilitating illness. While most individuals with severe RMSF experience complete recovery in 1–2 weeks following illness, roughly 1 in 4 patients may experience persistent or even permanent neurologic sequelae. We documented neurologic sequelae up to 11 years following acute disease. While most of the patients with LTS had 1 or more neurologic sequelae at discharge from their acute hospitalization, we also identified patients with LTS who were not appreciated to have deficits at time of discharge. This study provides the most comprehensive description of long-term neurologic sequelae following severe RMSF to date. While we are unable to attribute a specific cause of neurologic sequelae, we demonstrate that specific neurologic impairments, such as seizures, cognitive impairment, weakness, and damages to specific cranial nerves are consistently found in patients recovering from severe RMSF. Such results provide important and actionable information for patients, families, and healthcare providers in areas endemic for RMSF. Like patient recovery from other neurologic insults, such as stroke, patients who have been treated for severe RMSF would benefit from a complete neurologic examination at the time of discharge and potential use of neuroimaging studies to identify acute and subacute neurologic injury. Anticipation and early identification of neurologic injury can direct the use of rehabilitative services such as physical, behavioral, and occupational therapy in the months and years following acute disease to improve outcomes. Recognizing the full burden of RMSF can assist affected communities regarding planning, prevention, and outreach.

Notes

Acknowledgments. The authors acknowledge the contributions of the following individuals for their support during the medical record abstraction: Yari Torres-Mendoza, Stefanie Campbell, Carla Bezold, Cara Cherry, Edwin Rodriguez, Nicole Fowle, and Mary-Helen Wanat. The authors also thank Drs Sarah Coles, Mandy Boltz, and Christine Gill for their help during the neurologic exams, and especially thank the following individuals of the Public Health Nursing Departments and Community Health Representative programs in both communities for their support during the patient interviews including: David Bonito, J. T. Nashio, Jennette Declay, Laura Enos, Lillian Kavishe, DeAlva Larzelere, Carmelita Estenson, Nathanael Lemmon, Valerie Scott, Arlinda Newcomb, Lisa Mays, Gloria Cassa, Lolita Talkalai, and the other individuals who wished to remain anonymous. We thank Ken Komatsu, Arizona State Epidemiologist, for his review of the manuscript. Last, the authors recognize the contributions of Drs Rodrigo Villar and Christopher Paddock for their support during the design of this study.

Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention (CDC) or the Indian Health Service.

Financial support. This work was supported by the CDC.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.