SEARCH

SEARCH BY CITATION

Keywords:

  • anaphylaxis;
  • epidemiology

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Funding
  9. Conflicts of interest
  10. References

Background

There are very limited data characterizing the epidemiology of anaphylaxis from low- and middle-income country settings. We aimed to estimate the frequency of anaphylaxis admissions to hospitals in Istanbul.

Methods

We obtained data from all 45 hospitals in Istanbul over a 12-month period and used ICD-10 codes to extract data on those admitted with a recorded primary diagnosis of anaphylaxis. Because of concerns about possible under-coding, we undertook an additional analysis to identify patients admitted with two or more clinical codes for symptoms and/or signs suggestive of, but not coded as having, anaphylaxis.

Results

A total of 114 cases (79 people with anaphylaxis codes and 35 with symptoms and signs suggestive of anaphylaxis) were identified, giving an overall estimate of 1.95 cases per 100 000 person-years.

Conclusion

The novel two-stage identification approach employed suggests significant under-recording of anaphylaxis in those admitted to hospitals in Istanbul.

Recent studies on the epidemiology of anaphylaxis have found that the condition is more prevalent than once suspected and that it may be increasing in incidence [1-3]. The vast majority of these data, however, originate from economically developed countries [2-6]. The lack of data from low- and middle-income country settings therefore represents a major gap in our understanding the nature of this allergic emergency. In an attempt to begin to fill this evidence gap, we estimated the frequency of and patterns in anaphylaxis admissions to hospitals in Istanbul over a 12-month period.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Funding
  9. Conflicts of interest
  10. References

This study was approved by the Ethics Committee of Maltepe University Hospital.

We used the recently created regional database of The Health Directorate Network of Istanbul, which captures data on all hospital admissions from all state hospitals in the city to identify admissions over the last 12-month period for which comprehensive data were available, that is, November 1, 2010, through October 31, 2011.

We extracted data on all pediatric and adult admissions with a primary International Classification of Diseases Version 10 (ICD-10) coded diagnosis of anaphylaxis (Table 1). In addition, we extracted data on patients coded with two or more potentially relevant symptoms and signs of anaphylaxis [1] (Table 1).

Table 1. ICD-10-CM codes used to identify patients with coded and suspected anaphylaxis
ICD-10-CM codes for anaphylaxis
Anaphylactic shock due to adverse food reaction (T78.0)
Anaphylactic shock, unspecified (T78.2)
Anaphylactic shock due to serum (T80.5)
Anaphylactic shock due to adverse effect of correct drug or medicament properly administered (T88.6)
ICD-10-CM codes for symptoms and signs indicative of anaphylaxis
Skin-mucosal involvement: allergic urticaria (L50.0), idiopathic urticaria (L50.1), urticaria-unspecified (L50.9), generalized skin eruption due to drugs and medicaments (L27.0), generalized edema (R60.1).
Cardiovascular system: hypotension due to drugs (I95.2), idiopathic hypotension (I95.0) shock, not elsewhere classified (R57)
Respiratory system: wheezing (R06.2), asphyxia (R09.0), dyspnea (R06.0), respiratory arrest (R09.2)
Gastrointestinal system: nausea and vomiting (R11)
Nervous system: syncope and collapse (R55)

We analyzed data using the statistical package SPSS (SPSS for Windows, version 11.0, SPSS Inc., Chicago, IL, USA) and compared the differences between males and females in terms of case numbers, distribution of cases by age, and through the four seasons using chi-square and Fisher's exact tests, as appropriate. P-values less than 0.05 were considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Funding
  9. Conflicts of interest
  10. References

A total of 5 844 501 people (55.2% females) were admitted to these 45 hospitals during the study period. We identified 79 admissions with primary diagnosis codes specific for anaphylactic shock (i.e. 1.35 cases per 100 000 person-years) and a further 35 people with a combination of symptoms and/or signs suggestive of anaphylaxis (i.e. 0.6 cases per 100 000 person-years). This resulted in a total of 114 cases being identified who were likely to have been admitted for anaphylaxis (1.95 cases per 100 000 person-years, 95% CI: 1.3–3.77).

Most of the patients with an ICD code specific for anaphylaxis were older than 18 years (84%) and 57% were females (Table 2). In the second group, adults (85%) and females (74%) again represented the majority of cases identified (P < 0.05 for both). Anaphylaxis admissions took place predominantly in summer in both groups (P < 0.001 for both). Anaphylactic shock due to food (T78.0) was the most common specific code in both groups; in contrast, the code for anaphylactic shock due to serum (T80.5) was used in only one case.

Table 2. Key characteristics of the patients diagnosed as anaphylaxis with specific ICD-10 codes
 Total (n = 79)Females (n = 46)Males (n = 33)P-value
  1. a

    χ2 = 2.25.

  2. b

    χ2 = 27.46.

  3. c

    χ2 = 1.60.

Age (years)
<1010 (12.7%)4 (40%)6 (60%) 0.522a
11–183 (3.8%)2 (66.7%)1 (33.3%)
19–4034 (43.0%)19 (55.9%)15 (44.1%)
>4032 (40.5%)21 (65.6)11 (34.4%)
Season of diagnosis
Winter 16 (29.3%) 9 (56.2%)7 (43.8%) 
Spring 20 (25.3%) 12 (60%)8 (40%) 
Summer33 (41.7%)20 (60.6%)13 (39.4%) <0.001 b
Fall10 (12.7%)5 (50%)5 (50%) 
ICD-10 Codes
T78.033 (41.8%)19 (57.6%)14 (42.4%) 0.659c
T78.229 (36.7%)18 (62%)11 (38%)
T88.616 (20.3%)9 (56.2%)7 (42.8%)
T80.51 (1.2%)

There were no deaths in these 114 patients.

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Funding
  9. Conflicts of interest
  10. References

This study provides some of the first data on the epidemiology of anaphylaxis from low- and middle-income country settings. Our data suggest that hospital admission for anaphylaxis is a relatively uncommon recorded cause of hospitalization in Istanbul.

Epidemiological studies on anaphylaxis have yielded heterogeneous results, even in developed countries [2, 7, 8]. International estimates have suggested that the lifetime prevalence of anaphylaxis is about 0.05–2.0% [3], although data from both the USA and Europe suggest that the incidence of anaphylaxis is increasing [2, 3, 9]. Our overall estimates from Istanbul (1.9 cases per 100 000 person-years) are much lower than those emanating from the USA and Europe [2, 3, 9]. There may be four important reasons for this finding.

First, there is the possibility that patients experiencing anaphylaxis may be less likely to present to hospital than in other parts of the world. This is we believe, however, somewhat unlikely as all patients are able to attend hospital directly (i.e. without the need for referral from primary care), free of charge.

Second, knowledge of physicians and other healthcare workers on anaphylaxis in Turkey is not regularly updated and this may contribute to under-recognition and under-diagnosis; the lack of short, simple clinical and laboratory diagnostic criteria for the diagnosis of anaphylaxis may contribute to this under-diagnosis [1, 10]. This is because educating and training busy physicians in the use of complex diagnostic criteria such as those proposed by Sampson et al. [1] can prove particularly challenging in low-resource settings. Third, and related to, is the possibility that the heavy workload of physicians working in hospitals in Istanbul may contribute to the incorrect use of ICD codes for relatively uncommon conditions such as anaphylaxis. Building on the work of Harduar-Morano et al. [11], who reported a 58% increase in anaphylaxis cases by using an algorithm, we used a complementary approach to capture undiagnosed cases of anaphylaxis. Urticaria and/or angioedema are found in about 90% of patients with anaphylaxis and presence of these should be helpful in recognizing anaphylaxis [12, 13]. We were able to show that some of missed cases of anaphylaxis could be captured by investigating symptoms/signs involving other systems (i.e. respiratory, circulatory, digestive, and neurological) in all cases admitted to emergency rooms with skin and/or mucosal involvement.

Finally, anaphylaxis, especially due to food allergy, may in reality be less common than in the USA and northwest Europe [2, 12, 13]. Unfortunately, we do not have any comprehensive or reliable data on triggers for anaphylaxis in the Turkish population to enable meaningful comparison with other geographical settings.

It also needs to be borne in mind that most population-based studies have used ICD-9 codes to describe the epidemiology of anaphylaxis [14]. Comparing the results of these studies with our ICD-10-based analysis may therefore potentially be misleading. However, Wilson [15] used ICD-10 codes as the primary hospital admission diagnosis, thereby providing more comparable data: This analysis also found the incidence of anaphylaxis admissions to be much higher (i.e. 11.05 per 100 000 person-years) than that observed in our study.

The major strength of our work is that the estimates are derived from the investigation of the records of over five million people from all 45 hospitals serving a large urban population. Our data are therefore likely to provide a representative picture of the incidence of anaphylaxis admissions across Istanbul.

The main limitation of our study is that we interrogated routinely collected data and could not therefore confirm the accuracy of diagnoses. Another limitation is that we had no data about cases that did not present to hospital or those who developed anaphylaxis during the course of their admission. Our estimates are therefore likely to represent an under-estimate of the overall incidence of anaphylaxis in the city. Notwithstanding these limitations, we believe that this work represents the most comprehensive study into the epidemiology of anaphylaxis ever undertaken in a middle- and/or low-income country setting.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Funding
  9. Conflicts of interest
  10. References

The authors would like to thank Ms.Kubra Akbas for her valuable help in the collection of the data. We are grateful to the two anonymous reviewers for their constructive feedback on an earlier version of this manuscript.

Author contributions

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Funding
  9. Conflicts of interest
  10. References

Feyzullah Cetinkaya planned the study and led the writing of the manuscript. Ahmet Incioglu and Berrin Emre Karaman prepared the ethics submission and cleaned and analyzed the data. Suayip Birinci and Ali Ihsan Dokucu prepared the data in the Center of Health Directorate. Aziz Sheikh contributed to data interpretation and critically revising drafts of the manuscript.

Conflicts of interest

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Funding
  9. Conflicts of interest
  10. References

Aziz Sheikh has undertaken advisory work for ALK Abello, Lincoln Medical and Meda.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Funding
  9. Conflicts of interest
  10. References