Article Text

Download PDFPDF

Evaluation of a bedside immunotest to predict individual anti-tetanus seroprotection: a prospective concordance study of 1018 adults in an emergency department
  1. D Elkharrat1,
  2. M-J Sanson-Le-Pors2,
  3. L Arrouy1,
  4. A Beauchet3,
  5. F Benhamou1
  1. 1
    Department of Emergency Medicine, Université de Versailles St-Quentin, Assistance Publique-Hôpitaux de Paris, Paris, France
  2. 2
    Université de Paris 7, Assistance Publique-Hôpitaux de Paris, Laboratory of Microbiology and Virology, Lariboisière Hospital, Paris, France
  3. 3
    Department of Medical Information, Ambroise Paré Hospital, Boulogne-Billancourt, France
  1. Correspondence to Dr D Elkharrat, Service des Urgences Médico-chirurgicales, Hôpital Ambroise Paré, 9 avenue Charles de Gaulle, 92 100 Boulogne-Billancourt, France; david.elkharrat{at}apr.aphp.fr

Abstract

Background: Unscheduled tetanus prophylaxis (UTP) used in the emergency room (ER) in patients with wounds who are unaware of their vaccination history is erroneous in 40% of cases. Evaluation of bedside tetanus immunity with the Tétanos Quick Stick (TQS) test may improve UTP.

Objectives: To show that (1) a positive TQS result reflects immunity to tetanus; and (2) TQS is reproducible by ER workers.

Methods: In a prospective concordance study, immunity to tetanus of patients with wounds was assessed by two techniques: (1) TQS at the bedside, which detects specific tetanus antitoxins at concentrations ⩾0.2 IU/ml in whole blood or ⩾0.1 IU/ml in serum; (2) ELISA in the laboratory (threshold >0.1 IU/ml). The study comprised three groups: (A) healthcare personnel self-tested with the two techniques to determine the effect of training; (B) selected patients with wounds were double-tested with TQS by two healthcare providers whose readings were compared to test reproducibility; and (C) all patients with wounds aged ⩾15 years were consecutively included.

Results: Of 1018 individuals included, 60 were in group A, 50 were in group B and 908 were in group C. 403 patients who were not included were similar to those included for age, vaccination history and types of wounds. The reproducibility of the test was 98%. TQS sensitivity was 83.0%, specificity 97.5%, positive predictive value 99.6% and negative predictive value 42.9%.

Conclusions: TQS reliably predicts tetanus immunity and is reproducible by healthcare providers. Although it may not accurately discriminate between patients with ongoing and declining immunity, it is currently the most sensitive and specific tool for guiding tetanus prophylaxis and should be included in current guidelines on UTP.

Statistics from Altmetric.com

Current proof of individual anti-tetanus immunity (anti-TTI) is a serum tetanus antitoxin level (STAL) of >0.1 IU/ml when measured by enzyme-linked immunosorbent assay (ELISA).1 2 This technique, which requires 48–72 h to obtain results, is incompatible with proper unscheduled tetanus prophylaxis (UTP) in patients in the emergency room (ER) with accidental wounds. Consequently, UTP guidelines are designed by healthcare authorities worldwide.

These guidelines have two drawbacks: they are poorly adhered to by doctors3 4 and they rely on vaccination history. However, most patients seen in the ER do not carry certificates and are unaware of their immunity,3 so UTP is erroneous in over 40% of cases.3 4 Tetanus immunoglobulins (TIG) are overprescribed in 17–30% of patients despite the risks of inoculation.5 Conversely, UTP is suboptimal in 5–10% of patients who are generally poorly protected with the most tetanus-prone wounds.3 4

ER interventions have historically been successful in reducing unnecessary UTP,3 6 7 but not in reducing UTP underprescribings which entail a genuine risk. An immunochromatographic test, the Tétanos Quick Stick (TQS; Gamma, Belgium), has been marketed for its ability to assess anti-TTI at the bedside. If validated, it would improve UTP. Indeed, tetanus has been almost eradicated in Western Europe where its low annual prevalence, estimated at 0.5–1 per million inhabitants,8 9 10 is witness to the efficacy of vaccination. However, most cases of tetanus are severe11 and follow accidental wounds10 12 in unprotected patients.13 The annual prevalence of wounds in French ERs is estimated at 1.5 million,14 so it is important that reliable anti-TTI controls are carried out in such patients and that their protection is appropriately boosted.15 16

We conducted a prospective study in one of the busiest ERs in Paris in order to demonstrate that TQS might be an adequate tool for such a purpose if:

  • TQS and STAL measured by ELISA prove to be concordant—that is, a positive TQS results reflects anti-TTI and a negative result reflects the lack of it;

  • TQS is easily performed and reproduced by healthcare providers in ER working conditions.

Methods

TQS and STAL were blindly compared. TQS was read at the bedside by attending physicians and STAL measurements were determined by a laboratory technician who was unaware of the TQS outcome within 48–72 h of the ER visit. The cut-off value obtained with the ELISA was 0.1 IU/ml, as suggested by the TQS manufacturer. TQS was expected to be positive when STAL was >0.1 IU/ml and negative when ⩽0.1 IU/ml.

Techniques

TQS

TQS evaluates anti-TTI semi-quantitatively by detecting specific antitoxins in human serum, plasma or blood. The bar-shaped device (fig 1) has a sample well at one end and two windows, T (for test) and C (control), at the other. Its principle is based on the presence of tetanus toxoid conjugated with gold in a liquid phase at the well in a solid phase at the T window. When placed in the well, the sample runs to the windows. If antitoxins are present in the sample they form a complex with the conjugate which binds to the solid phase toxoid resulting in a pink stripe at the T window, indicating that the test is positive. A successful test requires a sample volume of 20 μl and the immediate addition of three drops of diluent to speed up migration. It is readable in 10 min. The C window is for quality control. Excess conjugate reaching the C window binds to a reagent, resulting in a pink stripe whether the sample contains antitoxins or not. A TQS unit is defective if the C window remains blank.

Figure 1

Tétanos Quick Stick (TQS): principle and operating mode.

The manufacturer of the TQS suggests a binary outcome—positive or negative. However, in some instances the T window may show a discontinuous or shadowy pink line. Such tests are considered equivocal.

STAL measurement by ELISA

STAL measurement by ELISA has been previously described.1 2 Blood samples from the patients are centrifuged and serum samples kept at −20°C until analysis. Three ranges are defined:

  • STAL ⩽0.1 IU/ml is non-protective; if known in real time it would indicate TIG for tetanus-prone wounds and complete vaccination in all patients.

  • STAL >0.5 IU/ml is considered protective and would require no UTP.

  • STAL >0.1 and ⩽0.5 IU/ml reflect an intermediate protection, usually on the decline, which would call for a toxoid injection.

Questionnaire

A questionnaire (available on request) was completed for all patients with wounds. One part was devoted to sociodemographic characteristics, immunisation history, type and mechanism of wound and its potential for tetanus. Another part featured definitions of wound contamination and adequate immunisation3 on which doctors were expected to base UTP. Although not a study objective, the use of the TQS result for UTP was encouraged.

Study groups

The study was conducted in three groups:

Group A: Training

Healthcare providers self-tested for TQS under supervision then trained their colleagues. They also tested for ELISA and self-administered the same interview as that given to patients, except for wounds.

Group B: Reproducibility

In each of 50 selected patients with wounds, two TQS were carried out by two healthcare providers using the same blood sample under supervision of the main investigator. Each proceeded to read his/her own test and then that of the other healthcare provider.

Group C

All consecutive patients with wounds of <24 h were included if they were ⩾15 years of age and signed consent (parental consent if aged <18) for the two techniques. Non-inclusion criteria were refusal to sign consent and/or submit to both tests, impaired consciousness or inability to answer questions, lesions requiring immediate surgery and prior inclusion in the study.

For all tested individuals, a drop of blood was obtained from the fingertip.

End points

The main end point was a near-zero rate of false positives (FPs), ie, TQS interpreted as positive when STAL was ⩽0.1 IU/ml. In addition, for TQS to be a reliable UTP guide, we aimed at a false negative (FN) rate of <1%. Finally, all ER healthcare givers should be able to perform and interpret TQS.

Statistical analysis

Based on a near-zero rate of FP and an FN rate of <1%, the total number of patients required was estimated at 1000. With regard to the first objective, sensitivity and specificity were estimated. Sensitivity was defined as the proportion of true positives (ie, individuals with positive TQS among those with STAL >0.1 IU/ml) and specificity was defined as the proportion of true negatives (ie, individuals with negative TQS among those with STAL ⩽0.1 IU/ml). FN and FP rates, positive and negative predictive values (PPV, NPV) and their 95% confidence intervals were calculated, the latter using the Confidence Interval Analysis software (BMJ, 2.1.2 version). Statistical analysis was performed with the program SAS (SAS Inc, Cary, North Carolina, USA).

Results

Of 1361 patients with wounds, 403 were not included (270 refused to sign consent, 75 refused to submit to TQS or STAL measurement, 25 had severe surgical lesions or consciousness impairment and 33 miscellaneous). The 1018 individuals included in the study were 60 healthcare personnel (group A) and 958 patients with wounds, 50 double-tested (group B) and 908 consecutively recruited (group C). Apart from the ER personnel, included and non-included patients were comparable (tables 1 and 2) except for the sex ratio (p = 0.016). For the 50 patients in group B, inter-observer reproducibility was 98%: all TQS readings in the first series of caregivers were identical and all but one of the second series.

Table 1

Demographic characteristics of included and non-included individuals

Table 2

Characteristics of wounds of included and non-included patients

TQS validation against the ELISA

All TQS performances for the three groups are shown in table 3. In groups A and B, TQS was technically supervised, which may explain their low FN rate and underlines the importance of personnel training.

Table 3

Sensitivity and specificity of TQS in individuals tested during three study phases

Although the population in group C truly reflects patients in the ER, most data will include 1018 individuals as they are similar to those of group C. Thus, global sensitivity was 83.0% (95% CI 80.4% to 85.3%), specificity 97.5% (95% CI 92.8% to 99.1%), PPV 99.6% (95% CI 98.9% to 99.9%) and NPV 42.9% (95% CI 37.1% to 48.9%), FN 17.0% and FP 2.5%. In other words, regardless of history, when TQS is positive the coexisting probability for STAL to be protective is 99.6%; when negative, the probability for STAL to be non-protective is 42.9%. The FP rate was 2.5% but the FN was higher than expected in all groups (table 3) and highest in medium range STAL (75%, 126/165) compared with 4% (12/293) when >0.5 IU/ml (table 4).

Table 4

TQS outcome according to STAL range

Prevalence of anti-tetanus protection

The prevalence of seroprotection (900/1018) by ELISA was estimated at 88.4% (95% CI 86.3% to 90.2%). It was considered intermediate/declining in 16.5% and solid in 71.9%.

Based on a positive TQS rate, the prevalence of anti-tetanus protection was 73.5% (95% CI 70.7% to 76.1%) in the 1018 individuals tested. Assuming that 17.0% of FN TQS reflect patients with intermediate/declining protection, anti-TTI by TQS (90.5%) and ELISA (88.4%) may be considered similar (table 4).

TQS and vaccination history

Vaccination statements were distinguished in assertive and negative/unsure patients (table 5). We defined history sensitivity as the proportion of seroprotected individuals stating they were vaccinated (527/900) and specificity as the proportion of non-vaccination statements in non-protected patients (102/118). Sensitivity was 58.5% (95% CI 55.3% to 61.7%), specificity 86.4% (95% CI 79.1% to 91.5%), NPV 21.5% (95% CI 18.5% to 25.4%) and PPV 97.0% (95% CI 95.3% to 98.2%). The contrast between the superiority of TQS sensitivity (83.0%) and NPV (42.9%) over those of history and the small difference between their respective specificities (97.5% and 86.4%) and PPV (99.5% and 97.0%) have been reported.17 Indeed, 13.5% of 118 individuals with STAL ⩽0.1 IU/ml and 67.4% of 626 with STAL >1 IU/ml stated they had complete and updated vaccination. Conversely, only 32.5% with STAL >1 IU/ml but 86.4% with levels ⩽0.1 IU/ml were unsure. In other words, unprotected patients are equally identified by TQS and immunisation history,3 4 but history alone may not identify those who mistakenly claim they have been vaccinated and are at risk of receiving no UTP. TQS may help in better identifying these patients.

Table 5

Tetanus immunisation according to vaccination history and STAL

TQS and unscheduled tetanus prophylaxis (UTP)

Local guidelines on UTP are aimed at reducing unnecessary TIG and boosting declining protection.3 UTP prescribed in 1350 of 1361 patients with wounds (data not shown), irrespective of inclusion or TQS administration, consisted of a booster in 75.8% of 190 stating no vaccination and no UTP in 89.5% of 563 claiming up-to-date vaccination. Of these 1350, 1149 patients received TQS. A booster was given in 23.3% of the total but only in 69% of 269 with a negative TQS and in 23.8% of 880 with a positive TQS. Had TQS outcome been applied, a booster would still have been given to 23.5% of 1149, but all with a negative test.

TQS and country of birth

The prevalence of anti-TTI in persons born in Africa, Asia, Latin America or Turkey and settled in Western Europe or the USA3 4 before the age of 15 was reported to be similar to that of people born in the host countries. Our study confirms these data (table 6) and shows a remarkable agreement between STAL and TQS in this regard.

Table 6

TQS and country of birth

Discussion

Characteristics of study population

In a previous study of patients with wounds in an urban setting,3 nearly 50% were born abroad; their wounds were not tetanus-prone in 88%. In the present study, 1361 consecutive patients with wounds of similar characteristics were enrolled (table 2).

Performance of the TQS

In practice, performing the TQS test is easy and painless and its outcome is available in 10 min. With proper training, TQS is reproducible by all ER professionals.

TQS was validated in three stages against ELISA, a reference for measuring STAL. Its performance, established on a large scale, is robust. Our data confirm the findings in other reports17 18 19 on the relevance of TQS in correctly predicting anti-TTI and directing UTP in individuals with wounds in the ER. A few points require discussion.

TQS false positive rate

TQS was falsely positive in three patients, the lowest reported in similar studies.17 18 19 None of the patients manifested clinical tetanus after 30 days, but FPs may appear unacceptable. Applied to the 1.5 million patients with wounds seen in ERs nationwide,14 it would translate into 38 000 cases annually. This concern may be mitigated in urban ERs where 88% of wounds are not prone to tetanus.3 Moreover, the FP rate may be kept very low by adequate training of professionals,17 20 as shown in the populations in groups A and B (table 3). Bearing in mind the possibility of FPs with TQS, we suggest aggressive UTP if tetanus-prone wounds are inaccessible to thorough cleaning in patients with risk factors for non-protection.3 4 This suggestion is even more justified as individual STAL measurements may vary from being protective long after vaccination21 to declining below the threshold within 10 years.22

The meaning of TQS positivity needs to be addressed. A more or less bright pink line may not discriminate between false and true positives or between solid and declining anti-TTI. However, the frequency of “protective” TQS was shown to rise when the stripe was entirely visible,23 usually reflecting STAL of >0.5 IU/ml. Equivocal TQS results are not infrequent,17 occurring in 3.5% of cases in this study. In our experience, a discontinuous or whitish strip which is difficult to visualise occurred consistently with intermediate/declining protection. Although TQS may be considered positive, we suggest boosting immunity, especially in wounds that are difficult to clean.

TQS false negative rate

Patients in group C are representative of wound management in the ER. The FN rate of 18.0% in this group is high, but does not represent a major drawback as it would only induce a toxoid surplus. Considering that the majority of FN tests reflect medium range STAL, the “excess” of toxoid at the time of the wound will help boost individual protection. There are two possible explanations for the high FN rate:17 18 the ELISA threshold chosen and the TQS technology, or both.

Although the manufacturer of the TQS suggests a STAL of ⩾0.1 IU/ml for TQS positivity, this cut-off value appears to be too low and poses a problem in real practice. Indeed, in US population-based STAL studies, the protective threshold was set at 0.15 IU/ml.24 25

Although the ELISA technique appears to be as sensitive and specific as passive haemagglutination, discrepancies occur between them, mostly for STAL <0.2 IU/ml. It may be so because haemagglutination identifies specific neutralising IgG, whereas ELISA quantifies various immunoglobulins including weakly neutralising IgM. Since TQS and haemagglutination are based on detecting specific neutralising IgG, it was suggested that TQS performances may benefit from STAL thresholds >0.15 IU/ml.17

We recalculated TQS performances in group C based on a cut-off of >0.15 IU/ml. The total of “unprotected” individuals rose from 109 to 142 and the FN rate decreased from 18.0% to 12.2% (data not shown). Also, sensitivity increased from 81.9.0% to 85.5% (95% CI 82.8% to 87.8%) with little change in specificity from 97.3% to 97.9% (95% CI 94.1% to 99.3%).

However, there remains an incompressible FN rate of about 10%, unaffected by threshold variations, which may be technology-dependent. To underline this point we used kept serum to randomly retest 40 initial negative TQS results of patients whose STAL was >0.5 IU/ml; TQS negativity was regularly confirmed.

Conclusions

The two study objectives—validation of TQS against ELISA and its reproducibility in ER working conditions—were achieved. Although it may not allow accurate distinction between intermediate/declining and ongoing anti-TTI, TQS is currently the most useful tool in diagnosing individual protection at the bedside and may be performed at ER triages.

In theory, a positive TQS result indicates anti-TTI and no requirement for UTP. However, a positive TQS result may have a pink line that is not clearly visible, which usually reflects declining protection; in these cases a toxoid booster is advised, and complete vaccination in elderly people. We advocate integration of the TQS test in current UTP guidelines (fig 2). It is best indicated in tetanus-prone wounds in patients with risk factors of declining or non-protection (essentially women aged >40 years and mean aged >60 years),4 14 17 24 25 26 27 birth in a developing country and settlement in a developed country after the age of 15 years3 17 19 and, arguably, elementary educational level or lower.4 24

Figure 2

Decisional algorithm for unscheduled tetanus prophylaxis in the emergency room. *TQS integration as a diagnostic tool to predict individual anti-tetanus protection.

With the appropriate use of TQS, UTP practices in the ER may show improved compliance and accuracy. Additional cost-effectiveness studies are required. Its input may also be considered in general practice, especially in institutionalised patients, and in advising travellers.23

Acknowledgments

The authors thank Mrs Nicole Cantaloube for her help in performing all the ELISA tests in the study; the healthcare personnel in the emergency room for their valuable collaboration in actively self-testing or in enrolling patients; and Nephrotek for providing the Tétanos Quick Stick and the immunosorbent assay used in the study.

REFERENCES

View Abstract

Footnotes

  • Competing interests None.

  • Ethics approval The protocol was approved by the local ethical committee.

  • Provenance and Peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Primary survey
    Darren Walter