Article Text
PDF
Abstract
Objectives To compare the proportions of Chlamydia trachomatis-positive specimens detected by Cobas Amplicor CT/NG (CA PCR) with C trachomatis positives in cell culture from 1999 to 2006 in order to estimate when the new variant of C trachomatis (nvCT) with a deletion in the cryptic plasmid (in the target region for CA PCR that resulted in false-negative results) emerged in Örebro County, Sweden.
Methods The annual number of specimens analysed using CA PCR in 1999–2006 ranged from 5077 to 11 622 and using cell culture (McCoy cells) from 5201 to 7425. Logistic regression was applied to evaluate the change in the proportion of C trachomatis-positive tests over the years between the two methods. The statistical interaction effect of year and method was estimated using both unadjusted and adjusted (age, gender and clinic) models.
Results From 2002, the proportion of C trachomatis-positive specimens identified using CA PCR decreased annually, whereas the proportion of culture-positive specimens increased annually. Logistic regression showed a statistically significant interaction effect between periods (1999–2006) and groups of specimens analysed using CA PCR or cell culture. A statistically significant association between the interaction of CA PCR/cell culture and period was observed in the unadjusted and adjusted models.
Conclusion This study indicates that in Örebro County, Sweden, nvCT was already present before 2002, that is, when the difference between the proportions of C trachomatis-positive specimens identified by CA PCR compared with cell culture-positive specimens began to show a statistically significant decline.
- Chlamydia trachomatis
- diagnostics
- epidemiology
- mutant
- nvCT
- statistics
- Sweden
Statistics from Altmetric.com
In 2006, a new variant of Chlamydia trachomatis (nvCT) was reported in Sweden.1 This variant had a 377 bp deletion in the cryptic plasmid that included the targets for the diagnostic systems from Roche Diagnostics and Abbott Laboratories, which at that time were widely used globally, and accordingly resulted in thousands of false-negative results.2 3 Both of the affected companies have now developed and CE mark-certified new dual-target assays capable of detecting nvCT.4
Several years ago in Sweden nearly all of the clinical laboratories replaced their cell culture methods and other established methods by nucleic acid amplification tests (NAAT) for the diagnosis of C trachomatis. However, in Örebro County (approximately 275 000 inhabitants) C trachomatis specimens were, at the time of the first report of nvCT, diagnosed using either Cobas Amplicor CT/NG (CA PCR; Roche Diagnostics, Pleasanton, California, USA) or a highly optimised and quality assured cell culture method (detecting nvCT), depending on the request from the submitting physician. Routine diagnostics of all specimens in Örebro County are performed at the Department of Laboratory Medicine, Section of Clinical Microbiology, Örebro University Hospital. During the past decade, the incidence of notified C trachomatis cases in Örebro County has been relatively similar to the national incidence (see figure 1, available online only), in which the number of cases increased from 157 per 100 000 inhabitants in 1997 to 367 per 100 000 in 2005. However, in 2006 the incidence decreased, from 367 to 357 cases per 100 000 inhabitants nationally and from 336 (year 2005) to 311 cases per 100 000 inhabitants in Örebro County.5 This decrease could subsequently be explained by the wide spread of nvCT, for example, in Örebro County the proportion of nvCT was 38% at the time of the first report of nvCT,5 approximately 30–35% during 2007 and mid-20084 and 24% at the beginning of 2009.6
The spread and proportions of nvCT have differed in the Swedish counties, which leads to the question: when did the mutant emerge and reach detectable levels in the different populations? Unfortunately, in 2006 few archival collections of C trachomatis samples were available at Swedish laboratories and it has not been possible to perform appropriate retrospective longitudinal studies to answer this question regarding the emergence of nvCT.
The aim of the present study was to compare the proportion of C trachomatis-positive specimens by the CA PCR method with the proportion of C trachomatis cell culture-positive specimens from 1999 to 2006 in order to estimate when nvCT first emerged in Örebro County, Sweden.
Materials and methods
Study population and laboratory methods
Urogenital or first void urine specimens were obtained from men and women attending the outpatient sexually transmitted disease (STD) clinic at Örebro University Hospital, the youth clinics, the welfare centres, the women's clinic, or other physisians in Örebro County and were sent for routine diagnosis of C trachomatis at the Örebro University Hospital. The diagnostic methods used were CA PCR (could not detect nvCT at that time) or cell culture (cycloheximide-treated McCoy cells), with subsequent identification of C trachomatis using fluorescein-labelled monoclonal antibodies (Phadebact Chlamydia IF Test, Bactus AB, Sweden) that can detect nvCT. The total number of specimens analysed using CA PCR for each year from 1999 to 2006 was 5077, 5178, 6877, 7811, 8702, 10 070, 11 111 and 11 622, and using cell culture 7425, 6666, 7133, 7311, 6446, 6148, 5533 and 5201, respectively. A low number of specimens (<1%) were excluded from the statistical analyses because of diagnostic problems such as cytotoxicity in the cell culture method and inhibited or equivocal results in the CA PCR. No specimen was analysed by both methods.
The percentage of C trachomatis-positive specimens from all clinics for each year (1999–2006) and method (CA PCR or cell culture) was calculated.
Statistical analyses
Logistic regression with a positive test (yes/no) as the outcome variable was used to evaluate the change in the proportion of positive tests over the years by estimating the statistical interaction effect between the two methods. The analysis was performed with variable year as a covariate to estimate the interaction as a linear effect. To visualise data the proportion of positive tests supplemented with 95% CI was plotted using the binomial distribution for each year and method in figure 2A and the difference in the proportion of positive tests between methods in figure 2B supplemented with 95% normal approximation CI. To analyse statistically in more detail in which year(s) the significant change in proportion occurred, years were categorised into 2-year intervals: 1999–2000, 2001–2, 2003–4 and 2005–6. The interaction effect was analysed between 1999–2000 and 2001–2, 2001–2 and 2003–4, and 2003–4 and 2005–6. All analyses were performed unadjusted and adjusted. In the adjusted analysis age, gender and clinics were included as covariates, to compensate for potential confounding effects. The effect parameters are expressed as OR supplemented with 95% CI.
(A) Percentage of Chlamydia trachomatis-positive tests using Cobas Amplicor CT/NG (CA PCR) and cell culture from 1999 to 2006 supplemented with 95% CI based on the binomial distribution. (B) Difference in the percentage of C trachomatis-positive tests using CA PCR and cell culture from 1999 to 2006 supplemented with 95% CI.
Results
The distributions of patients examined for C trachomatis using CA PCR or cell culture in 2-year intervals from 1999 to 2006 are presented in table 1. Among the specimens analysed using CA PCR, 51% were from patients attending the STD clinic and the youth clinic (high-risk group), in total, 59% were women (median age 20–24 years). In contrast, among the specimens analysed using the cell culture method only 33% were from patients attending these clinics, whereas 50% were from welfare centres (low-risk group), in total, 95% were women (median age 20–24 years).
The distribution of patients examined for C trachomatis using CA PCR or cell culture (in percentages) by clinics, gender and age in 2-year intervals from 1999 to 2006
Estimating only the specimens analysed using CA PCR, the proportion of positive tests decreased from 1999 to 2006, by an average of 3.5% per year in the unadjusted model and 3.9% per year in the adjusted model. In contrast, estimating only the specimens examined using cell culture, the proportion of positive tests increased during the same period by an average of 2.3% per year in the unadjusted model and 3.7% per year in the adjusted model (figure 2A). When analysing year as a linear effect from 1999 to 2006 the logistic regression analysis showed a statistically significant interaction effect between methods (CA PCR and cell culture) and years (1999–2006) on the proportion of positive tests both in the unadjusted (p<0.001) and adjusted model (p<0.001) (table 2).
Interaction of a C trachomatis-positive test using CA PCR or cell culture by year (1999–2006) as a linear effect
Furthermore, the difference between the proportions of positive tests using the two divergent methods plotted each year (figure 2B) showed a visual change in the distribution of positive tests between methods beginning in 2002–3.
Comparing the years 1999–2000 with 2001–2, no significant statistical interaction effect between methods and period was found in either the unadjusted (p=0.997) or adjusted model (p=0.945) (table 3). However, the interaction effect was significant when comparing the intervals 2001–2 and 2003–4, both in the unadjusted (p=0.030) and adjusted model (p=0.036). Moreover, this significant interaction of methods and period was observed when comparing the intervals 2003–4 and 2005–6 (unadjusted p=0.010 and adjusted p=0.001). This finding showed that the difference between the proportions of C trachomatis-positive specimens by CA PCR compared with cell culture-positive specimens began to decline statistically significantly between the periods 2001–2 and 2003–4.
The interaction of a C trachomatis-positive test analysed using CA PCR or cell culture and year intervals: 2001–2 compared with 1999–2000, 2003–4 compared with 2001–2 and 2005–6 compared with 2003–4
Discussion
The present study demonstrated that nvCT was already prevalent at non-negligible levels in 2002–3 in Örebro County (table 3, figure 2B), that is when the difference between the proportions of C trachomatis-positive specimens by CA PCR and in cell culture began to decline statistically significantly. All specimens from patients attending the outpatient STD clinic at Örebro University Hospital, the youth clinics, the welfare centres, the women's clinic, or other physicians in Örebro County and analysed for C trachomatis between 1999 and 2006 were included. The specimens were diagnosed using either CA PCR (could not detect nvCT at that time) or cell culture (detecting nvCT) and the C trachomatis positivity rates were compared statistically for the 8-year period (1999–2006). The estimation was based on a comparison of the positivity proportions using two diagnostic methods (CA PCR and cell culture), with nvCT seemingly present before 2002 in Örebro County (figure 2A, B). The proportion of C trachomatis-positive specimens diagnosed by CA PCR was higher, because approximately 50% of the specimens were from patients attending the STD clinic and youth clinics (ie, high-risk group). The corresponding figure analysed by cell culture was approximately 30%. No major alterations of these methods or any preventive interventions in the population studied were performed during the years examined that could explain the findings.
The incidence of notified C trachomatis cases in Örebro County (see figure 1, available online only) also indicates that the increase in the incidence, which has been stable since the mid-1990s, already showed a reduction in 2002, which is in full concordance with our statistical analyses. Furthermore, already in 1999–2000 a suspected culture-positive mutant, which was positive by conventional ompA (omp1) gene PCR but repeatedly negative in CA PCR, was identified in Örebro County.7 This specimen was determined as genotype E (ie, as the present nvCT), but 47% of the specimens in that study were genotype E. At that time, the isolate was considered to be an exceedingly rare plasmid-free C trachomatis strain.8 9 However, it cannot now be excluded that this was a nvCT sample already present in 1999–2000 in Örebro County, although it was not until 2002–3 that the spread of nvCT had reached an impact on the C trachomatis population, which was possible to identify with the statistical analyses used in the present study. We have previously described that all the examined nvCT isolates in Örebro County have been of genotype E, with an identical ompA sequence and displayed an identical and unique multilocus sequence typing profile.5 10 This profile was also found in arbitrarily selected nvCT isolates (n=48) from four countries,11 which strongly indicates a clonal spread of nvCT.
In another Swedish study a trend analysis was performed in 2004 to 2006 on national data from the Swedish Institute for Infectious Disease Control of the number of genital chlamydial cases detected and the number of C trachomatis tests performed.11 This analysis showed a decline in the proportion of positive chlamydia tests in the counties that used the Abbott or Roche methods from 2005, whereas the positivity rates in the counties using the BD ProbeTec ET (Becton Dickinson, Franklin Lakes, New Jersey, USA) tests did not change. The discrepancies between the present study and the previous national study11 regarding the year nvCT emerged may be that in the national study only a few years (2004–6) and the first 6 months of each year were studied. Another difference is the nationwide coverage of the previous study, which included all the counties of Sweden that significantly differed in population characteristics, C trachomatis incidence and prevalence, diagnostic facilities and preventive and other interventions performed, all of which may affect the positivity rates of the C trachomatis diagnostics. Nevertheless, the county-specific differences in coverage in testing of the target population were adjusted for.
Since 2006 nvCT has been detected in high numbers throughout Sweden. However, the proportions of nvCT now seem to converge in the Swedish counties after the selective diagnostic advantage for nvCT has disappeared in the Roche/Abbott counties.6 This observation is also supported by the unaltered biological fitness of nvCT, which was identified in a recent study comparing the nvCT genome with all available C trachomatis genome sequences, and also comprehensively investigating the biological properties of nvCT.12 Astonishingly, although many studies have been performed worldwide, for the most part only sporadic nvCT cases have been identified outside the Nordic countries.13 14 Nevertheless, knowledge of the presence and prevalence of nvCT in other countries is currently limited for the following reasons: few recent studies have been conducted, many European laboratories still cannot detect nvCT because they are using CA PCR or inappropriate in-house PCR, and laboratories that can detect nvCT do not know it because no nvCT-specific or other distinguishing NAAT are used.15 16 Accordingly, ideally all laboratories should use NAAT that can detect nvCT, because a wider geographical spread of nvCT cannot be excluded. Furthermore, greater emphasis should be given to the importance of regular local, national and international surveillance for unexplained declines in the incidence of C trachomatis. Finally, recommendations for diagnostics and participation in appropriate external quality assessment schemes need to be emphasised.
In conclusion, we were able to compare the C trachomatis positivity rates from 1999 to 2006 using two methods (Cobas Amplicor CT/NG PCR and cell culture on McCoy cells) within the same laboratory in Örebro County, Sweden. Based on our statistical analyses, nvCT seems to have emerged before 2002, that is when the difference in the proportions of C trachomatis-positive specimens by Cobas Amplicor PCR compared with cell culture-positive specimens began to show a statistically significant decline.
Key messages
A statistical estimation was performed to indicate when nvCT emerged in a defined area of Sweden.
From 1999 to 2006 the proportion of C trachomatis-positive specimens detected by CA PCR was compared with C trachomatis positives using the cell culture method.
At that time (ie, 1999–2006) CA PCR could not detect nvCT, whereas the cell culture method was able to detect nvCT.
The statistical estimation in the present study indicates that nvCT was present before 2002 in Örebro County, Sweden.
References
Footnotes
Funding This study was funded by the Örebro Medical Research Foundation, Örebro University Hospital, Örebro, Sweden.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Whistlestop tour