| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
 |
ABSTRACT |
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Gastrointestinal symptoms are experienced during a short period of time in the acute phase of the disease. This phase has been defined by some authors as a period ranging anywhere from 2 days to 1 week.2,4,5 Cardiac symptoms such as bradycardia are not present in all patients and have been proposed as an indicator of the severity of the disease.4,6,7 These symptoms also seem to disappear and reappear in the first stage of the disease.8 In contrast, the chronology of neurologic and diffuse manifestations is not as precise and seems to follow a different pattern.4,9–11 For example, some case reports suggested that neurologic and diffuse symptoms appeared later than gastrointestinal troubles and that they may persist for weeks, months, and even years after onset.4,12,13 Although many outbreaks of ciguatera have recently been reported, epidemiologic studies on the severity or duration of ciguatera disease are rare.6,14,15 Thus, no consensus exists on the exact duration of these symptoms, and some authors, such as Pearn (2001; see Ref. 4), deplore the lack of a survival curve of symptoms of ciguatera suitable for patients and the management of the disease.
Although ciguatera disease presents a low fatality rate (<0.1%), it presents a large gradient of severity spreading from intermittent paresthesia in the mouth to severe polyneuropathy, hence greatly limiting quality of life.16–19 Severity of the disease in case reports is mainly based on the type of symptoms experienced by the patient. Nevertheless, some epidemiologic studies have suggested that severity varies according to age, gender, and history of previous intoxications.6,11,14,15,20 Moreover, others have proposed that recurrence of symptoms could have been triggered by fish, protein, or alcohol consumption, physical activity, sexual intercourse, or other individual susceptibility factors.6,14,15,21,22 To our knowledge, no systematic study, on such a large population sample, has been performed on the frequency or on the determinants of persistence of neurologic symptoms of ciguatera disease. In this context, this prospective study aims to address the clinical evolution of ciguatera poisoning and identify associated factors in French Polynesian adults.
|
SUBJECTS AND METHODS |
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Consent forms and questionnaires were sent by health care workers to the research team in Papeete at the Institut Louis Malardé. Each diagnosis of ciguatera was confirmed by a clinical investigator upon reception of the first epidemiologic questionnaire. A ciguatera case was defined as the acute onset of neurologic symptoms after (<12 hours) having consumed a good-tasting local reef fish. Typically acute symptoms included paresthesia and troubles in the perception of cold, with or without gastrointestinal symptoms such as vomiting and diarrhea. Exclusion criteria were any one of the following: 1) age < 15 years, 2) pregnancy, 3) leprosies or sensory/motor neurologic diseases, 4) ciguatera diagnosis > 7 days after the onset of the disease. According to these criteria, we excluded 2 children < 15 years old, 1 patient who did not have neurologic symptoms, and 17 patients who had been recruited in the study even though they did not respect the 7-day period defined in the protocol as the onset period. The final sample retained in our analysis included 182 patients.
Parameters recorded at T1.
The symptoms scale was built from 26 self-reported symptoms, present in the first questionnaire, except for cardiac variables, which were obtained by objective measures. We grouped symptoms and signs in 4 classes as follows: (a) gastrointestinal symptoms (nausea, diarrhea, and vomiting); (b) neurologic symptoms (paresthesia, trouble with cold perception, headache, dental pain, metallic taste, vertigo, instability, lack of coordination, itching, hypersalivation); (c) cardiac symptoms [bradycardia (pulse 
60 bpm), tachycardia (
90 bpm), hypotension (systolic 100 mm of Hg); hypertension (systolic/diastolic above 140 /90 mm of Hg)]; (d) systemic symptoms (asthenia, myalgia, arthralgia, dysuria, fever, cervical rigidity, respiratory distress, chills, sweating). An index of the variability of symptoms was then created based on these classes of symptoms. This index, categorized into 3 levels ("low," "medium," and "high"), took into account if patients experienced 1, 2, 3, or all classes of symptoms, respectively. The higher the index, the more extended the disease. Distribution of the population according to the index severity scale is presented in Table 1
.
|
Parameters recorded at T2. Thirteen symptoms were recorded and scored on a severity scale based on Lange’s25 modified symptoms check list rating scale that was developed specifically for quantifying severity and monitoring therapy responses in ciguatera poisoning. All symptoms were recorded and scored according to their frequency of persistence during the week preceding individual interviews as follows: "Never," 0; "Once a week," 1; "Twice a week," 2; "3 or 4 times during the week," 3; "5 times and more," 4. Scores varied between 0 and 52. We then aggregated them in a scale in 4 levels, such as 0 (asymptomatic patient), 1 (mildly symptomatic, score < 12), 2 (moderately symptomatic, score = 12–26), and 3 (severely symptomatic, score > 26), which was used as our main outcome.
In the same questionnaire, patients quantified their tobacco and alcohol consumption and completed a specific dietary questionnaire on type of proteins consumed, such as meat (beef, chicken, pork, corned beef, and processed meat), fish (lagoon fish, ocean fish, and seafood), and dairy products (milk, yogurt, and ice cream).
Statistical analysis. Summary data are expressed as means and standard deviations for continuous variables, or proportions for categorical variables. Skewed variables were normalized by logarithm transformation and back transformed for easy interpretation.
At T1, univariate ordinal logistic regressions were used to estimate the odds ratio (OR) of the index of variability of symptoms according to factors recorded in the questionnaire. The models were built according to proportional odds assumption. The logs obtained were cumulated over the levels of response. The appropriateness and the goodness of fit for the proportional odds models were examined by testing the score and the Pearson statistics. The preselection criterion of potential confounders was an association with a dependent variable with a P value < 0.20.26 We considered a confounding effect if the OR of the severity of the disease in the model showed a change of 10% or more after adjustments with the potential confounder.
To evaluate the persistence of symptoms 15 days after the onset, we treated data as repeated-measure outcomes using the generalized estimating equation (GEE) approach adapted to ordinal outcomes.27,28 This method provided standard errors adjusted by multiple observations per person using an independent correlation structure. This analysis provided cumulative logs that focused on the comparison of lower to higher values of the symptom-persistence scale according to factors recorded in the follow-up adjusted for time period. In this analysis, the potential confounding effects of other variables were also examined with preselection criteria identical to those criteria presented above.
Except for the selection of potential confounding variables in the multivariate analysis, by convention, we considered a P value < 0.05 to be statistically significant. All statistical treatments were performed using SAS software, version 8.02 (SAS Institute, Inc., Cary, NC).
|
RESULTS |
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
presents characteristics of the 182 patients retained in the analysis according to the index of variability of symptoms. At the time of the onset, the 3 groups of patients appeared similar with respect to 2 factors studied: delay before consultation and quantity of fish ingested (all P values < 0.01). This was confirmed by results (see Table 2), where significant associations were detected between the index of variability of symptoms and fish consumption-related factors. These associations remained significant after adjustments except for part of fish consumed (Table 2).
|
, except for itching (64% at 15 days after onset), which remained stable. For this analysis, we used data only from patients who completed both interviews (N = 113).
|
|
DISCUSSION |
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Even if a common description of ciguatera cases exists, we noticed, as proposed by others,2 that the number of symptoms experienced at onset and the type of symptoms simultaneously experienced vary considerably from one patient to another. Our results highlighted a predominant influence of factors related to fish consumption, which play a key role in the quantity and type of ciguatoxin ingested.29 The dose–response relationship observed was previously noticed by other researchers.14,15 Moreover, we found that people who consumed organ (head, liver, or gonad) were more likely to develop more severe ciguatera than those who ate only filet. Although this failed to reach significance after adjustment, it is in accordance with other observations and biologic studies that indicated that the organs (head, liver, or gonad) are the tissues showing the highest concentrations of the toxin in fish.12,30 This distribution of toxins in suspected fish is supported by a decrease in the index of variability of symptoms according to an increase of fish intake. Organs were quantified in the study as the smallest possible portion compared with other parts of the fish.
In earlier studies, authors had reported associations between individual factors such as age, sex, lifestyle, past ciguatera event, and severity of disease.6,11,14 In this study, we found none of these associations.
Our analysis of the persistence of symptoms of ciguatera showed that no symptoms had totally disappeared 15 days after onset. These results reinforced previous observations9 that suggested that 
20% of patients had symptoms persisting for several months. Moreover, we noticed that symptoms most frequently reported after 2 weeks were mainly systemic symptoms such as asthenia, myalgia, and headaches. A similar association of symptoms had been previously observed by Pearn, who described an evolution of ciguatera disease that could be linked to chronic fatigue syndrome.31 Our results also corroborated with remarks from Shnorf and colleagues who proposed in 2002 pruritus as a late clinical manifestation of ciguatera disease.32 Apart from its tardy appearance, pruritus also appears in our study as a long-lasting symptom with the highest prevalence among those patients who have symptoms remaining 15 days after onset. This could explain the common name "gratte" (itching) given to ciguatera disease in most of the South Pacific. The late appearance of pruritus, and its long duration, allow us to propose itching as the characteristic sign of the chronic stage of ciguatera disease, just as cold allodynia is for the acute onset of the disease.
In our study, only 2 factors were clearly associated with the persistence of symptoms. The first is the index of variability of symptoms defined at onset. Patients with a higher score experienced more symptoms which were more frequent than did people with a lower score. The second factor is tobacco consumption, which may impede healing. This association with smoking was observed for the first time, and further studies are required to clarify the intervention of tobacco as a possible risk factor of the persistence of symptoms. Moreover, some authors14,33,34 have noticed worsening symptoms related to alcohol, protein, as well as fish consumption. In our results, none of these factors seemed to be significantly associated with the persistence of symptoms.
Nevertheless, it is important to keep in mind that this study suffers from a major loss of patients during the follow-up. Indeed, the original design of the study called for 4 interviews (at onset and then 15 days, 2 months, and 6 months after the onset). Unfortunately, the attrition rate decreased from 38% at 15 days, to 64% 2 months later, and reached 75% at 6 months after the onset. The main hypotheses that could explain these important losses to follow-up are length of the study (length of questionnaire and physical examination) and nature of symptoms experienced (i.e., not sufficiently incapacitating to require medical assistance). However, inspection of data showed that patients who quit the study were similar at the onset to patients who decided to pursue. Consequently, we believe that patients who remained in the study were not those who were more severely ill, nor did they introduce an important overestimation of the prevalence of chronic symptoms. Taking into account these limits, it appears interesting to note that among 30 people who answered the third questionnaire (i.e., 2 months after the onset), 12 of them still experienced symptoms. Although these last observations could not support a general conclusion on the duration of ciguatera disease, they at least suggest that symptoms may persist in some patients who were not severely intoxicated at the onset. Indeed, none of the patients recruited in this study had respiratory distress, were unconscious, or had heart failure. Consequently, our population may be qualified as mildly intoxicated compared with other case reports previously published. In the same way, it is important to keep in mind that the index of variability of symptoms constructed here is based on subjective signs and cannot be extrapolated to a severity index of the disease. However, it provided interesting information on the number of physiologic systems altered by the disease.
In light of these results, it would be interesting, in future investigations, to corroborate our observations and go into detail, using objective measures of symptoms experienced by patients, to clarify the magnitude of symptoms reported. At a prevention level, our results corroborate and reinforce previous conclusions that have contributed to the development of fish consumption advisories on avoidance of organ (head, gonad, and liver) consumption of high-risk species.
Received May 26, 2005. Accepted for publication August 15, 2007.
Acknowledgments: The authors thank the Centers for Disease Control and Prevention (CDC) and Pan American Health Organization (PAHO) for financial support of the study. We are grateful to local general practitioners for their participation to this study, to patients for their cooperation, and to ILM’s research nurses for their help. We also thank Ms. Susan Gingras and Mr. Elhadji A. Laouan Sidi for their expert advice on statistical analysis.
* Address correspondence to Eric Dewailly, Public Health Research Unit, Laval University Medical Research Center, CHUL, CHUQ, Québec, Canada. E-mail: eric.dewailly{at}crchul.ulaval.ca 
Authors’ addresses: Eric Dewailly and Marie-Ludivine Chateau-Degat, Public Health Research Unit, Laval University Hospital Research Centre (CHUQ), 945 Avenue Wolfe, 2nd floor, Sainte-Foy, Québec G1V 5B3, Canada, Telephone: +1 (418) 650-5115, ext. 5240, Fax: +1 (418) 654-3132. Marie-Odile Huin-Blondey, Direction de la Santé Service de Protection Maternelle, BP 527 Papeete, Tahiti, French Polynesia, Telephone/Fax: +689-81-29-92. Mireille Chinain and Taiana Darius, Laboratoire Microalgues Toxiques, Institut Louis Malardé, BP 30, 98713 Papeete, Tahiti, French Polynesia, Telephone: +689-41-64-58, Fax: +689-43-15-90. Anne-Marie Legrand and René Chansin, Institut Louis Malardé, BP 30, 98713 Papeete, Tahiti, French Polynesia, Telephone: +689-41-64-60, Fax: +689-43-15-90. Ngoc Lam Nguyen, Centre de Consultation Médicale et d’Investigation Epidémiologique, Institut Louis Malardé, BP 30, 98713 Papeete, Tahiti, French Polynesia, Telephone: +689-41-64-53, Fax: +689-43-15-90. François Laudon, Direction de la Santé Service de Protection Maternelle, BP5287- 98716, Pirea, Tahiti, Telephone: +689-50-91-75, Fax: +689-3-67-35.
|
REFERENCES |
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
