SEARCH

SEARCH BY CITATION

Keywords:

  • anaphylaxis;
  • IgE antibodies;
  • neuromuscular blocking agents;
  • pholcodine withdrawal

Abstract

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Statement of contribution
  8. Conflict of interest
  9. References

To cite this article: Florvaag E, Johansson SGO, Irgens Å, de Pater GH. IgE-sensitization to the cough suppressant pholcodine and the effects of its withdrawal from the Norwegian market. Allergy 2011; 66: 955–960.

Abstract

Background:  IgE-mediated anaphylaxis to neuromuscular blocking agents (NMBA), frequent in Norway, was proposed to be caused by exposure to pholcodine (PHO) carrying the allergenic quarternary ammonium ion epitope. Consequently, the PHO-containing drug was withdrawn from the market in March 2007.

Objective:  Describe the effects of withdrawal of PHO on IgE, IgE-antibodies and reported frequencies of anaphylaxis to NMBAs.

Methods:  Three hundred sera from supposedly allergic patients sampled yearly through 2006 to 2010 were analysed for IgE antibodies to PHO, suxamethonium (SUX) and morphine (MOR). Furthermore, IgE and preliminary reports from the Norwegian Network for Anaphylaxis under Anaesthesia (NARA) were monitored.

Results:  PHO exposure was associated with IgE sensitization to PHO, MOR and SUX. However, after withdrawal, within 1 year, antibody prevalences to PHO and SUX fell significantly from 11.0% to 5.0% and from 3.7% to 0.7%, respectively. At 3 years, SUX had fallen to 0.3%, PHO to 2.7% and MOR to 1.3%. By 2 years, the prevalence of elevated IgE was significantly reduced. After 3 years, the incidence of reported suspected anaesthetic anaphylaxis fell significantly, both the total number, the reactions related to NMBAs and those with IgE antibodies to SUX.

Conclusions:  Withdrawing of PHO lowered significantly within 1–2 years levels of IgE and IgE antibodies to PHO, MOR and SUX, and, within 3 years, the frequency of NMBA suspected anaphylaxis. The results strengthen the PHO hypothesis considerably and equally the need to question the existence of cough depressants containing PHO.

During general anaesthesia, adverse reactions in the form of anaphylaxis have been increasing in frequency during the past decades (1). Although rare in numbers, they constitute a threat to patient safety and a challenge to the diagnostic and therapeutic skills of anaesthetists. Mortality rates as high as 3.5% and 10% have been reported, and crucial for the outcome is early recognition and optimal treatment including adrenaline (2–4). The reactions mainly appear to be IgE-mediated, and most countries report neuromuscular blocking agents (NMBAs) as major causes, suxamethonium (SUX) being the prominent culprit. Further intriguing features, besides their notorious unpredictability and the absence of valid risk factors, are that reactions usually take place without prior exposure to anaesthetic drugs and that national incidence reports vary beyond reasonable explanations.

Against this background, in 2004, studies were set up to compare the two closely related countries Norway and Sweden as in Norway anaphylactic reactions to NMBAs were about ten times more frequent than in Sweden (5). In 300 sera from the allergy laboratories at the University Hospitals in Bergen (Norway) and Stockholm (Sweden), IgE antibodies to morphine (MOR) (6) and SUX were not detected in the Swedish, but in 10% and 3.7%, respectively, of the Norwegian samples (7). The exposure to environmental chemicals containing the quarternary ammonium ion epitope (QAI), postulated as the binding site for IgE antibodies towards NMBAs (8), was also examined. However, no major differences were found between the two countries, except for the wide use of a cough mixture (Tuxi®; Weifa AS, Oslo, Norway) containing pholcodine (PHO) in Norway, but not in Sweden (7).

PHO exposure of individuals IgE-sensitized to PHO showed extraordinary effects on IgE synthesis. A week’s intake of 1/3 of the recommended daily dose of cough syrup resulted in 50 and 100 times elevated serum levels of both IgE and IgE antibodies to PHO, MOR and SUX (9). After stop of exposure, normalization of serum levels seemed to follow two courses, either back to pre-exposure levels within a few months, or in some individuals, elevated levels were maintained for several years (9, 10).

The observed discrepancies between Norway and Sweden and the responses to PHO exposure provided the basis for the ‘pholcodine hypothesis’ (11). High consumption of PHO through cough mixtures was related to increased prevalences of IgE antibodies to PHO, MOR and SUX and higher incidences of IgE-mediated anaphylactic reactions to NMBAs.

The aims of this interventional study were to describe the consequences of PHO exposure on IgE and IgE antibodies to PHO, MOR and SUX in Norway. Further, to monitor the effects of PHO withdrawal on prevalences of sensitization and incidences of anaphylactic reactions to NMBAs.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Statement of contribution
  8. Conflict of interest
  9. References

Samples

From the Allergy Section of the Laboratory of Clinical Biochemistry (Accredited according to NS-EN ISO 15189), Haukeland University Hospital, Bergen, Norway, three nested groups of anonymized serum samples (A, M and H) were collected from volumes superfluous after the requested IgE analyses had been performed. The samples were submitted to the laboratory from general practitioners (GPs) and specialists in the region to evaluate suspected allergy. Group A (‘Allergics’) contained 300 consecutive samples collected without knowledge of clinical data or IgE levels but with a residual volume of ≥350 μl. ‘Allergics’ Group M (Medium) consisted of 100 sera with IgE levels between 1000 and 5000 kU/l. In ‘Allergics’ group H (High), the intention was to collect 50 sera with IgE levels >5000 kU/l. Sampling was performed during the months of April to June each year.

In addition to the above, all analyses of IgE performed during the 12 months preceding withdrawal of Tuxi® in March 2007, termed ‘Before’ (n = 24096), were compared to the results obtained during the following 12-month periods of 2007 ‘1 year’ (n = 24129), 2008 ‘2 years’ (n = 25806) and 2009 ‘3 years (n = 26491). The data were statistically evaluated for changes in the prevalences of IgE levels above 120 and 5000 kU/l, respectively.

Immunoglobulin analyses

IgE and IgE-antibodies to PHO, MOR and SUX were determined by the ImmunoCAP® system (Phadia AB, Uppsala, Sweden) according to the instructions of the manufacturer. Sensitization prevalences were given as percentages of sera with IgE antibody levels of ≥0.35 kUA/l. For comparisons of allergenic epitopes, values down to 0.1 kUA/l were used.

Anaphylaxis reporting

The Norwegian Network for Anaphylaxis during Anaesthesia (NARA, Haukeland University Hospital, Bergen) based on collaboration between anaesthetists and allergologists at the major Norwegian University Hospitals receives, in a spontaneous reporting system, standardized clinical reaction data and sera collected at the time of the suspected reactions. The Allergy Section of the Laboratory of Clinical Biochemistry, Haukeland University Hospital, analyses the serum samples for tryptase, IgE and IgE antibodies (Phadia AB) to a panel of relevant allergens including SUX.

The study was approved by the Regional Committee for Medical Research Ethics in Western Norway.

Statistical evaluations

Variations in percentage of samples with elevated (>120 kU/l) and high (>5000 kU/l) levels of IgE by year were calculated by negative binomial regression with log link to test relative risk (RR). ‘Before’ was used as reference group. Relative risk was adjusted for gender, age and month of analysis. Linear trend for adjusted RR by year was tested. Time trends of IgE-sensitization to PHO, MOR and SUX and reported suspected anaphylactic reactions were calculated by negative binomial regression with log link. For tables comprising cells with few numbers, linear trend was also calculated by exact methods in cross-tables. In the analyses, the total number of general anaesthesias per year is set to 200 000. Association between PHO and MOR, SUX and IgE were tested by linear regression analysis. All statistical analysis was performed by Statistical Package for the Social Science 17, (SPSS Inc 2008, Chicago, Illinois, USA).

Results

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Statement of contribution
  8. Conflict of interest
  9. References

The impact of PHO exposure on IgE-sensitization in the three subpopulations of ‘Allergics’ when compared with blood donors is summarized in Table 1. Percentages of sera with IgE antibody levels above cut-off (≥0.35 kUA/l) to MOR, PHO and SUX are considerable among ‘Allergics’. Most notably, in the two allergic populations, medium (M) and high (H), as many as 1/3 and 3/4, respectively, were IgE-sensitized to PHO. IgE-sensitization to SUX, i.e. to the QAI, rose from 3.7% among general ‘Allergics’ to 12.0% and 30.6%, respectively, in the M and H groups. Sensitization is also clearly present among blood donors, where 6% carry IgE antibodies to PHO and 0.4% to SUX. Although there is a clear association between IgE sensitization to MOR and SUX in all groups, between 46% and 93% of positive still do not show antibody binding to QAI/SUX.

Table 1.   Prevalences of IgE sensitization (≥0.35 kUA/l) to PHO, MOR and SUX in blood donors, ‘allergics’ in general (Group A) and ‘allergics’ selected by having medium (Group M; IgE 1000–5000 kU/l) and high (Group H; IgE >5000 kU/l) serum IgE levels. Analyses were performed during the final years before the PHO-containing cough mixture Tuxi® was withdrawn from the Norwegian market. The results are expressed in absolute numbers with the respective percentages in parenthesis
Sample groupsNo.MOR (%)PHO (%)SUX (%)
  1. MOR, morphine; PHO, pholcodine; SUX, suxamethonium.

  2. *The results for sample groups marked with asterisk are taken from Ref. 7.

Blood donors*50025 (5.0)30 (6.0) 2 (0.4)
Allergics group A*30130 (10.0)33 (11.0)11 (3.7)
Allergics group M10031 (31.0)33 (33.0)12 (12.0)
Allergics group H4937 (75.5)36 (73.5)15 (30.6)

The principle associations between IgE antibodies to PHO, MOR and SUX are exemplified by group M data. The plotted individual antibody levels show a significant, linear relationship between PHO and MOR (gradient of regression 1.35; P < 0.001) (Fig. 1A). In contrast, the relationship between PHO and SUX (Fig. 1B) and IgE was not significant. Of the 62 PHO pos sera, 39 were SUX negative, i.e. 63%.

image

Figure 1.  IgE antibody levels to pholcodine (PHO) given as kUA/L and plotted against morphine (MOR) (A) and suxamethonium (SUX) (B). The gradient of the regression line of PHO-MOR measurements was 1.35 (P < 0.001), while for PHO-SUX, no significant gradient was seen. The individually plotted antibody levels are taken from the ‘Medium Allergics’ serum group M (n = 100) with total IgE 1000–5000 kU/l (Table 1). Please note that if more than one serum was negative (<0.1 kUA/l) for SUX only one dot is shown for each PHO-value.

Download figure to PowerPoint

Removal of Tuxi® from the Norwegian market resulted in significant decreases in the prevalences of IgE-sensitization to PHO, MOR and SUX in the ‘Allergics’ group A (Table 2). Already one year after withdrawal, the percentages of IgE antibodies to PHO and SUX were reduced from 11.0% to 5.0% and from 3.7% to 0.7%, respectively. At three years, SUX sensitization had continued to fall to 0.3%, PHO to 2.7% and MOR to 1.3%. All the changes were statistically significant (P < 0.001).

Table 2.   Prevalences of IgE sensitization (≥0.35 kUA/l) to PHO, MOR and SUX in ‘allergics’ (Group A, n = 300) sampled yearly up to 3 years after withdrawal of the PHO-containing cough syrup Tuxi® and compared to the before withdrawal data from Table 1. The results are expressed as absolute numbers and, in parenthesis, percentages of positive samples
 BeforeYears after Tuxi® withdrawalTrendP-valueP-value Linear-by linear
123
  1. MOR, morphine; PHO, pholcodine; SUX, suxamethonium; nt, not tested.

  2. *Exact methods.

PHO33 (11.0)15 (5.0)17 (5.7)8 (2.7)−0.418<0.001<0.001*
MOR30 (10.0)nt8 (2.7)4 (1.3)−0.667<0.001<0.001*
SUX11 (3.7)2 (0.7)1 (0.3)1 (0.3)−1.0660.002<0.001*

Comparing the number of yearly analyses of IgE (n approximately 25000) performed at the routine allergy laboratory over the years from 2006 to 2009 shows that Tuxi® also affected the profile of this variable. A significant decrease in the prevalences of elevated IgE was seen for the 120 kU/L cut point at 2 and 3 years after withdrawal (Table 3). Similarly, a significant fall (not shown) in the numbers of sera with IgE ≥5000 kU/l was found, from 83 (0.34%) before to 58 (0.24%) and 44 (0.17%) at 2 and 3 years, respectively, after withdrawal (P = 0.001 for the falling trend in RR, adjusted for age, gender and month of analysis).

Table 3.   Occurrences of elevated IgE (>120 kU/l) in per cent of the total number of yearly analyses performed at the routine allergy laboratory and related in time to the withdrawal of Tuxi®: 2006 – before (n = 24096), 2007 – after 1 year (n = 24129), 2008 – after 2 years (n = 25806) and 2009 – after 3 years (n = 26491). 2006 was used as reference year for statistical analysis
 Per cent IgE >120 kU/lCrude RR95% CIAdjusted RR*95% CI
  1. *Relative risk (RR) was adjusted for gender, age and month of analysis.

2006 before25.31 1 
2007 after 1 year26.31.041.01–1.101.041.01–1.07
2008 after 2 years24.10.950.90–0.980.950.92–0.98
2009 after 3 years21.50.920.77–0.850.880.84–0.91
Trend   0.810.001

The incidence of NARA reports of suspected anaphylactic reactions in the operating theatre, both the total number and the number related to the use of NMBA, has decreased significantly at 3 years after withdrawal (Table 4). A similar and significant trend was seen for the total number of sera with IgE antibodies to SUX at the time of reaction, decreasing in 2009 to three and in the first 6 months of 2010 to two cases (further NARA data to be published).

Table 4.   Linear trend of reported suspected anaphylactic reactions during general anaesthesia from the NARA registered yearly through 2005–2010. Shown are the total numbers of reported reactions (Total) and the number of reactions related to the use of NMBAs (NMBA). In addition, the number of sera with IgE antibodies to SUX (≥0.35 kUA/l) measured at the time of reaction is listed (IgE SUX). For statistical analysis, the total number of general anaesthesias per year is set to 200 000
 200520062007200820092010 first 6 monthTrendP-valueP-value Linear by linear
  1. NARA, Norwegian Network for Anaphylaxis under Anaesthesia; NMBAs, neuromuscular blocking agents; SUX, suxamethonium.

  2. *Exact method.

Total948981885325−0.116<0.001<0.001
NMBA576256663418−0.0890.0200.018
IgE SUX1118121532−0.1980.0210.017*

Discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Statement of contribution
  8. Conflict of interest
  9. References

This study has shown that PHO consumption, in Norway through the widespread use of the cough mixture Tuxi®, had a major impact on IgE synthesis. High prevalences of IgE-sensitization to PHO was seen especially in allergics, but were also quite noticeable in blood donors. Withdrawing the drug from the market in March 2007, led, already after one year, to a significant fall in prevalences of IgE antibodies to PHO, MOR and SUX and after 3 years to a significant drop in reported numbers of suspected anaphylactic reactions to NMBAs in line with observations from Sweden in the 1980’s (12).

The results lend strong support to the PHO hypothesis (11, 13) and carry a number of clinical implications. It has been shown in previous publications (7, 9, 10, 14), and again underlined in this study, that the stimulation of IgE synthesis by PHO is associated with the production of IgE antibodies to the QAI, present bi-valently on NMBAs like SUX. This finding allows the option of preventing IgE-mediated anaphylactic reactions caused by NMBAs simply through avoiding exposure to PHO in cough mixtures, an intervention that presently seems to prove efficient in Norway. However, it still remains to be seen whether similar interventions would do equally so for other high PHO consuming countries like France, UK and Australia. Also, the recent suggestion that other, yet unknown, substances might possess sensitizing potentials similar to PHO (14) needs further investigations.

While Tuxi® was on the market in Norway, it was estimated by the manufacturer that about 40% of the population were taking the drug. That about 5–10% of the population became IgE-sensitized to PHO is remarkable, indicating that PHO is a potent sensitizer. We have previously suggested, based on our own in-hospital data, that being IgE-sensitized to SUX (i.e. QAI) when receiving a general anaesthesia including SUX increases the risk of anaphylaxis from one in 5200 anaesthetics to about one in 20 (13, 15). So, being IgE-sensitized from exposure to PHO represents a considerable risk factor for anaphylaxis in the operating theatre. The British Association of Anaesthetists reported 55 documented cases of anaphylaxis per year with a fatality rate as high as 10% (16). However, comparing PHO consumption and prevalences of anaphylaxis to NMBA in France, Norway and UK, one gets the impression that there is an underreporting in the latter country.

Because, however, not every SUX-sensitized individual reacts with anaphylaxis upon intravenous SUX administration, other factors besides the presence of the initiating IgE-antibody probably play a role. One such factor could be that the QAI hapten is exposed differently in vitro and in vivo and that the pathogenic epitope is at least partially dependent on a so far not characterized carrier. Like, for instance, the allergenic epitope of Patent Blue V, a colour injected to trace sentinel lymph nodes in e.g. breast cancer, which is completely dependent on the hapten/carrier conjugate (17).

Another factor that could influence clinical reactivity is that during induction of anaesthesia, a number of QAI carrying drugs including MOR or MOR derivatives are administered. Because these are monovalent, they could inhibit, but not initiate, the reaction of IgE antibodies to QAI present on SUX (7) and other NMBAs. Like for other hapten blocking (18), the effect as to the inhibition of IgE-mediated anaphylaxis would probably in part depend on dose and timing relative to the administration of NMBA.

The polyclonal effect of PHO on IgE synthesis may further offer explanations for two clinical situations familiar at an allergy outpatient clinic. One is referral of patients with elevated serum levels of ‘total IgE’, which could not be explained by clinical allergy or other causes like atopic eczema, parasitic infestations or a number of immune pathological conditions. The other is referral for unspecific clinical symptoms where the GP has found serum IgE antibodies, mostly low levels, to a variety of airway and/or food allergens with uncertain clinical significance. Based on our experiences, as a (former) high PHO consuming country, a principle approach in such cases is to explore the history for use of PHO-containing cough medicines (evidence of exposure) and to analyse for IgE antibodies to PHO (evidence of sensitization) and SUX (possible risk factor for IgE-mediated anaphylaxis to NMBA). According to the results in Table 1, one-third of IgE concentrations between 1000 and 5000 kU/l and three-fourths of IgE levels higher than 5000 kU/l were related to PHO sensitization. However, in the latter group, because some serum samples had a very high IgE level, a certain degree of nonspecific binding can not be excluded.

In diagnosed cases, treatment includes terminating PHO exposure and, eventually, monitoring normalization of serum IgE and IgE antibody levels. When exposure stops, IgE levels seemingly declines in two definite courses as indicated in the present and other studies (9, 10). In most cases, normalization will take place within a year or two. However, in some individuals, the decline towards normal can be more prolonged maybe attributed to boostering by ‘house-hold chemicals’ carrying the QAI epitope. It should also be remembered that PHO was rather quietly taken off the market, and thus, many individuals may have, for a considerable time, continued taking the Tuxi® they had at home.

Another challenge in PHO-sensitized individuals is related to anti-IgE treatment of severe allergic asthma. The decision whether this treatment is indicated and the calculation of the correct dosage are based upon the levels of serum IgE. In addition, the notion is that these levels remain fairly stable over time. In individuals IgE-sensitized to PHO, however, the serum IgE concentration is very high and may vary 100-fold within a week or two (9, 10).

After the detection of the substituted ammonium ion as the major allergenic determinant in NMBAs (8), MOR has been suggested for the screening of IgE antibodies to NMBAs (6). We have previously cautioned against this strategy because of specificity items. At least in a population exposed and sensitized like the Norwegian, it would yield a high number of false-positive results. The data from the present study (Table 1) support this view by showing that among the three allergic serum populations A, M and H, about half and more of the sera positive to MOR/PHO were negative to SUX. Further, in serum from blood donors, 93% of those positive to MOR/PHO were negative to SUX (7). One could argue that this discrepancy in part may be explained by a low sensitivity of the used SUX ImmunoCAP (1). A more likely explanation could be our previous finding that MOR and PHO have two non-cross-reacting IgE binding epitopes, one of which is QAI (7). Thus, an immunoassay for IgE antibodies to MOR picks up IgE antibodies with two different specificities: those to the QAI epitope shared with SUX and the NMBAs, and those to the MOR/PHO epitope which is not present on NMBAs. It is possible that the latter epitope is more immunogenic than QAI. The fact that PHO is monovalent for its two non-cross-reacting allergenic epitopes might explain the very rare incidence of severe allergic reactions to cough syrups containing PHO (19, 20).

In conclusion, PHO exposure has shown strong effects on IgE synthesis in Norwegian allergic and blood donor populations. The resulting production of IgE antibodies to PHO, MOR and SUX, and thus to the QAI epitope, offers, according to the PHO hypothesis, an explanation for the frequent reporting of anaphylactic reactions during anaesthesia. However, terminating sales of the PHO-containing cough mixture showed remarkably effective, leading to significantly less sensitization and fewer reported reactions within 1–3 years. The results strengthen the hypothesis and, more so, the need to question the use of cough suppressants containing PHO.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Statement of contribution
  8. Conflict of interest
  9. References

The authors acknowledge BMA Ingjerd Møller and BMA Marit Økland, Laboratory of Clinical Biochemistry, Haukeland University Hospital for systematic collection of serum samples and analysis of IgE and IgE antibodies, and for helping to retrieve the stored IgE results for 2005–2010. Further, Prof. Anne Berit Guttormsen, Department of Anaesthesiology and Intensive Care, Haukeland University Hospital, Bergen, Norway is thanked for preparing the preliminary NARA data.

Statement of contribution

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Statement of contribution
  8. Conflict of interest
  9. References

Prof. E. Florvaag and Prof. S.G.O. Johansson: Chest physician and clinical immunologist. Responsible for the study design, contacts with authorities including EC and contacts with Drug Authorities. Å. Irgens: Statistician. Statistical analyses and evaluations. Dr G. H. de Pater: Anaesthesiologist. Has examined and evaluated the preliminary data made available from NARA.

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Statement of contribution
  8. Conflict of interest
  9. References