Fax: (713) 794-4662
Meta–analyses of the association between Chlamydia psittaci and ocular adnexal lymphoma and the response of ocular adnexal lymphoma to antibiotics
Version of Record online: 26 JUN 2007
Copyright © 2007 American Cancer Society
Volume 110, Issue 4, pages 809–815, 15 August 2007
How to Cite
Husain, A., Roberts, D., Pro, B., McLaughlin, P. and Esmaeli, B. (2007), Meta–analyses of the association between Chlamydia psittaci and ocular adnexal lymphoma and the response of ocular adnexal lymphoma to antibiotics. Cancer, 110: 809–815. doi: 10.1002/cncr.22843
- Issue online: 2 AUG 2007
- Version of Record online: 26 JUN 2007
- Manuscript Accepted: 11 APR 2007
- Manuscript Revised: 27 MAR 2007
- Manuscript Received: 31 JAN 2007
- Chlamydia psittaci;
- ocular adnexal lymphoma;
- mucosa-associated lymphoid tissue (MALT) lymphoma;
There are conflicting reports regarding the association between Chlamydia psittaci (Cps) and ocular adnexal lymphoma (OAL) and the efficacy of antibiotics for OAL. In the current study, the authors attempted to clarify the association between Cps and OAL and the efficacy of antibiotics for OAL.
Two meta‒analyses were conducted. One focused on the association between Cps and OAL across geographic regions and among different studies. The other was a meta‒analysis of the response of OAL to antibiotic treatment.
The authors identified 11 studies of Cps prevalence that included 458 cases of OAL from 10 different countries. Four studies regarding the efficacy of oral antibiotics to treat OAL were found. One hundred four of the 458 OAL specimens (23%) and 87 of the 346 mucosa‒associated lymphoid tissue (MALT) lymphoma specimens (25%) were found to be positive for Cps. Ninety-four of the 104 Cps‒positive OAL specimens (90%) came from 3 of the 11 studies. There was wide variation noted between geographic regions and even between studies from the same geographic region with regard to the rate of Cps positivity. The 4 studies concerning the efficacy of antibiotics for OAL were from Italy, Austria, Taiwan, and the U.S. and included 42 patients. Twenty patients (48%) achieved some response (complete response in 8 patients, partial response in 8 patients, and minimal response in 4 patients). Twenty patients also had stable disease, and 2 patients progressed during antibiotic therapy. Objective documentation of response (radiographs or clinical slit-lamp photographs) was available for only 3 of the 42 patients. Seven additional patients developed disease recurrence after their initial response or stable disease after antibiotic therapy; 6 of these cases of disease recurrence occurred during the first 12 months of follow‒up.
The findings of the current study suggest a striking variability in the association between Cps and OAL across geographic regions and even between studies from the same geographic regions. The overall rate of Cps positivity in our meta‒analysis (23%) was much lower than that reported in the original report. The current study findings also suggest that antibiotics have variable efficacy against OAL. Future prospective trials with standard objective response criteria and a longer follow‒up period would be necessary to evaluate the role of antibiotics in the treatment of OAL further. Cancer 2007; 110:809–15. © 2007 American Cancer Society.
Ocular adnexal lymphoma (OAL) is a subtype of non-Hodgkin lymphoma; the most common histologic type of OAL is mucosa‒associated lymphoid tissue (MALT) lymphoma.1 MALT lymphoma, a low‒grade lymphoma, accounts for 60% to 80% of all cases of OAL.2, 3 Some recent studies suggest a possible increase in the incidence of MALT lymphoma in the last decade.1
Much attention has been focused on determining whether OAL is caused by an autoimmune disorder, chronic antigenic stimulation, or both. Prior studies have linked gastric MALT lymphoma with Helicobacter pylori,4 cutaneous MALT lymphoma with Borrelia burgdorferi,5 and splenic marginal zone lymphoma with hepatitis C.6 Recently, in a study from Italy, Ferreri et al. reported an association between Chlamydia psittaci (Cps) and OAL.7 However, several other studies from other regions of the world have failed to demonstrate such an association.8–16
Because of the evidence that MALT lymphoma may be caused by chronic infection, some authors have investigated the efficacy of antibiotics in the treatment of OAL, both in patients with confirmed bacterial infection and in patients with unknown infection status. Reports published to date regarding the efficacy of antibiotic treatment for OAL have presented conflicting findings.17–21
We conducted 2 meta‒analyses to shed more light on the association between Cps and OAL and the efficacy of oral antibiotics for the treatment of OAL.
MATERIALS AND METHODS
The MEDLINE and PubMed databases were searched for publications regarding the association between Cps and OAL and for publications concerning the efficacy of antibiotics for the treatment of OAL. The literature search spanned the period between April 2004 (when what to our knowledge was the first report of the association between Cps and OAL was published) and October 2006. For each study, the following details were recorded: the geographic region in which the patients with OAL were treated, the age of the patients, patient gender, the histologic classification of the lymphoma, the stage of lymphoma, the method of DNA detection of Cps, and Cps status. For the meta‒analysis of the efficacy of antibiotic treatment, the duration of antibiotic treatment and the response to treatment were recorded. Tumor response was classified as complete response (eradication of all clinical evidence of tumor), partial response (≥50% reduction in the tumor), minimal response (<50% reduction in the tumor), stable disease (no change), or progressive disease (growth of the tumor by at least 25%).22
Association Between Cps and OAL
We found 11 studies published between April 2004 and October 2006 regarding the association between Cps and OAL (Table 1). These included specimens from 458 patients from 10 different countries and 4 different continents; 1 study included patients from > 1 country. The breakdown of studies and patients by country was as follows: 3 studies from the U.S. (20% of patients),8, 15, 16 3 studies from Italy (19% of patients),7, 16, 19 1 study from China (12% of patients),16 2 studies from the Netherlands (9% of patients),11, 16 1 study from the U.K. (9% of patients),16 2 studies from Japan (8% of patients),13, 14 1 study from Korea (7% of patients),10 1 study from Germany (6% of patients),16 1 study from Cuba (6% of patients),9 and 1 study from France (3% of patients).12
|Study||Location||No. of patients with OAL||Median age, y||Male/Female ratio||Histologic subtype (No. of patients)||Site (No. of patients)||Cps positivity rate, No. of patients (%)|
|Ferreri et al., 20047||Italy||40||66||0.7||MALT (24)||Lacrimal (11)||OAL: 32/40 OAL (80%)|
|Diffuse large B-cell lymphoma (5)||Conj (8)||MALT: 21/24 MALT (88%)|
|Other lymphoma (11)||Orbit (21)|
|Gracia et al., 20069||Cuba||26||MALT (19)||OAL: 2/26 (8%)|
|MALT: 2/19 (11%)|
|You et al., 200510||Korea||33||42||1.1||MALT (33)||26/33 (78%)|
|Rosado et al., 20068||U.S.-Miami||57||65||0.48||MALT (49)||Lacrimal (13)||0/57 (0%)|
|Non-MALT (8)||Conj (16)|
|Mulder et al., 200611||Netherlands||19||MALT (19)||0/19 (0%)|
|de Cremoux et al., 200612||France||16||Conj (10)||1/16 (6%)|
|Daibata et al., 200613||Japan||21||56*||1.0||MALT (18)||Lacrimal (4)||0/21 (0%)|
|Non-MALT (3)||Conj (6)|
|Liu et al., 200614||Japan||17||50||0.55||MALT (12)||0/17 (0%)|
|Diffuse large B-cell lymphoma (4)|
|Small lymphocytic lymphoma (1)|
|Vargas et al., 200615||U.S.-New York||11||78||1.75||MALT (7)||Lacrimal (1)||0/11 (0%)|
|Non-MALT (4)||Conj (4)|
|Chanudet et al., 200616||195||62||1.3||MALT (142)||Conj (114)||31/142 MALT(22%)|
|Non-MZL (53)||Orbit (88)||5/53 non-MZL (9%)|
|U.K.||40||64||0.9||MALT (33)||MALT: U.K.—4/33 (12%) +1 non-MZL|
|Germany||28||58||1.3||MALT (19)||MALT: Germany—9/19 (47%)|
|Netherlands||24||62||0.6||MALT (21)||MALT: Neth—6/21 (29%) +2 non-MZL|
|Italy||21||61||0.5||MALT (15)||MALT: Italy—2/15 (13%)|
|South China||57||60||2.6||MALT (37)||MALT: China-4/37(11%) +2 non-MZL|
|Eastern U.S.||25||69||1.3||MALT (17)||MALT: U.S.—6/17 (35%)|
|Ferreri et al., 200619||Italy||27†||56||0.42||Lacrimal/Orbit (13)||11/27 (41%)|
In 8 of the 11 studies, the method of DNA detection of Cps was modeled after the method of Madico et al.23 The first step in their technique was to use multiplex touchdown enzyme time‒release polymerase chain reaction (PCR) on DNA extracted from formalin‒fixed, paraffin‒embedded tissues.23 This step detects 3 species of Chlamydia (Chlamydia trachomatis, Chlamydia pneumonia, and Chlamydia psittaci) simultaneously. Next, CPS100 and CPS101 primers were used to produce a 111‒base pair amplicon. These primers amplified highly conserved regions encoding the 16S rRNA gene, which is conserved in Chlamydia species, and the 16S‒23S rRNA spacer, which varies in different Chlamydia species and is unique for Cps.14 The 3 studies that did not use the technique of Madico et al.23 only performed the multiplex touchdown enzyme time‒release PCR using DNA extracted from formalin‒fixed, paraffin‒embedded tissues, which detects Chlamydia (all 3 species), but did not perform the second step, which would provide more specific information for Cps.
The age range and gender distribution of the patients were available for 7 of the 11 studies. In these 7 studies, there were 202 men and 179 women, and the median age was 59 years (range, 8‒95 years).
Of the 458 patients in the 11 studies, 346 (76%) had MALT lymphoma; 9 patients (2%) had diffuse large B‒cell lymphoma; 1 patient had small lymphocytic lymphoma; and 79 patients (17%) had disease identified as nonmarginal zone lymphoma, non‒MALT, “other,” or not specified. The location of OAL was specified in 382 patients (83%): 167 patients (36%) had orbital tumors; 172 patients (38%) had conjunctival tumors, 29 patients (6%) had lacrimal gland tumors, and 1 patient (0.2%) had an eyelid tumor. In 13 patients (3%), lacrimal gland and orbital involvement were included as a single entity. In 95 patients (21%), the exact location of the OAL was not specified. Only 2 studies (including 22% of the patients) included staging of the OAL. Approximately 80% of the patients in these studies had stage I disease and approximately 20% had stage IV disease.
The proportions of patients who tested positive for Cps are shown in Table 1. One hundred four of the 458 patients with OAL (23%) were positive for Cps. Of the 346 patients with ocular adnexal MALT lymphoma, 87 (25%) were positive for Cps. Approximately 90% of the OAL specimens that were positive for Cps DNA (94 of 104 specimens) came from just 3 of the 11 studies included in this meta‒analysis, and the rate of Cps positivity among the specimens in those studies was 35%. In stark contrast, the rate of Cps positivity in the other 8 studies was only 7% (14 of 194 specimens).
There was much geographic variation in Cps positivity (Table 2). Rates of Cps positivity varied even between studies from the same region. For example, the rate was 80% in the first study published by Ferreri et al.,7 41% in a subsequent study by Ferreri et al.,19 and only 10% in another study by Chanudet et al.16 The rate was 24% in 1 of the U.S. series16 and 0% in the other 2 American reports,8, 15 and was 33% in 1 Dutch series16 and 0% in the other.10
|Geographic location||No. of studies included||Cps positivity|
|Korea||You et al., 200510||78%|
|Italy||Ferreri et al., 20047; Chanudet et al., 200616; and Ferreri et al., 200619||49%|
|Germany||Chanudet et al., 200616||32%|
|Holland||Mulder et al., 200611 and Chanudet et al., 200616||19%|
|England||Chanudet et al., 200616||13%|
|China||Chanudet et al., 200616||11%|
|U.S.||Rosado et al., 20068; Vargas et al., 200615; and Chanudet et al., 200616*||6.5%|
|France||de Cremoux et al., 200612||6%|
|Cuba||Gracia et al., 20069||4%|
|Japan||Daibata et al., 200613 and Liu et al., 200614||0%|
Efficacy of Antibiotics for the Treatment of OAL
We found 4 studies published between April 2004 and October 2006 regarding the efficacy of antibiotics for the treatment of OAL (Table 3).17–21 These studies included 42 patients: 27 patients (64%) from Italy, 11 patients (26%) from Austria, 3 patients (7%) from the U.S., and 1 patient (2%) from Taiwan. There were 16 men and 26 women. The patients ranged in age from 18 to 87 years (median, 60 years). Forty‒one patients were treated with doxycycline at a dose of 200 mg per day for at least 3 weeks. One patient in the study by Abramson et al.17 was treated with a PrevPak (lansoprazole at a dose of 30 mg, amoxicillin at a dose of 500 mg, and clarithromycin at a dose of 500 mg) twice a day for 14 days. The median follow‒up time after antibiotic treatment was 11.5 months.
|Study||Location||No. of patients treated||Median Age, y||Male/Female ratio||Cps status||Treatment||Treatment stage (No. of patients)||Objective response, No. of patients||Median follow-up after treatment|
|Ferreri et al., 200619||Italy||27||56||0.42||11 positive cases per multiplex PCR 16 -||100 mg doxy twice daily × 3 wk||1st line (15)||†CR, 6 (22%)||14 mo|
|2nd line (12)||PR, 7 (26%)|
|MR, 3 (11%)|
|SD, 9 (33%)|
|PD, 2 (7%)|
|Grunberger et al., 200621||Austria||11||63||0.83||Not performed||200 mg doxy daily × 3 wk||1st line (11)||SD, 11 (100%)||9 mo|
|Yeung et al., 200420||Taiwan||1||18||Male||Positive serology||100 mg doxy twice daily × 3 wk||2nd line||MR||6 mo|
|Abramson et al., 200517||U.S.||3||75||2.0||Negative serology in all 3||100 mg doxy daily × 4 wk (1 patient received PrevPak‡ × 14 d)||NA||CR, 2 (66%)||28 mo|
|PR, 1 (33%)|
In the study by Ferreri et al.,19Cps status was verified via touchdown enzyme time‒release PCR, and all 11 of the 27 patients (41%) were found to be Cps positive.19 In the study by Yeung et al., the 1 patient was negative for Cps by serology.20 The study by Grunberger et al.21 did not include testing for Cps status. In the study by Abramson et al., all 3 patients were negative for Cps by serology.17 Of the 19 patients who were negative for Cps (16 by PCR and 3 by serology), 9 (47%) were reported to have achieved a response to antibiotics (4 patients achieved a complete response and 5 achieved a partial response).
Eight (19%) of the 42 patients with OAL who were treated with antibiotics achieved a complete response, 8 patients (19%) achieved a partial response, 4 patients (10%) achieved a minimal response, 20 patients (48%) had stable disease, and 2 patients (4%) developed progressive disease. Overall, 22 patients (52%) experienced either no improvement or a worsening of disease, whereas 20 patients (48%) experienced some improvement. Seven additional patients in the study by Ferrari et al. developed disease recurrence after an initial response or initial stable disease after antibiotic treatment, with 86% of the recurrences occurring during the first 12 months of follow‒up.19
The distribution of responders across geographic areas was as follows: all those patients with complete and partial responses were reported in the Italian (81%) and U.S. studies (19%). The 1 patient in the Taiwanese study achieved a minimal response at 6 months after antibiotic treatment. All 11 patients in the Austrian study had stable disease at 9 months.
Our findings in these 2 meta‒analyses suggest that there is significant variation in the prevalence of Cps DNA in OAL specimens and that this prevalence varies by geographic region and is not as high as reported in the early studies of this association. The current study findings further suggest that oral antibiotics are not effective against the majority of cases of OAL and indicate that the majority of reports in the literature of a positive response of OAL to antibiotics are not accompanied by objective documentations of response.
Overall, approximately 23% of OAL specimens had PCR evidence of Cps DNA. There was a great deal of variability between geographic regions and even between studies from the same geographic region with regard to the prevalence of Cps positivity. The highest Cps positivity rate was reported by Ferrari et al. in their first article, which reported an overall Cps positivity rate of 80%.7 In a more recent article, the same authors reported a much lower rate of Cps positivity (41%).19 The Korean study by You et al. reported a Cps positivity rate of 78%.10 However, 2 other studies of Asian patients have reported much lower rates. Daibata et al.13 reported a rate of 0% in Japanese patients, and Chanudet et al.16 reported a rate of 11% among Chinese patients. Variability in the methodology used to detect Cps DNA may be a possible explanation for the wide variability observed in Cps positivity rates.
Another explanation for the wide variability in Cps positivity rates could be that the presence of Cps DNA is not etiologically related to the occurrence of OAL and thus the variability observed is simply a reflection of the variability in exposure to Cps on the conjunctival surface or other ocular adnexal structures. Exposure to infected birds and cats is a risk factor for psittacosis, a lung infection caused by Cps. In the initial study by Ferreri et al., contact with household animals was reported in 13 patients,7 all of whom were positive for Cps. In the same study, 14 patients with OAL also had chronic conjunctivitis; 12 of these patients had PCR evidence of Cps DNA. It is difficult to estimate the prevalence of Cps DNA in ocular adnexal tissue in healthy adults; arriving at such an estimate would require random sampling of ocular adnexal tissue collected from large populations of healthy adults, and to our knowledge this has not been done to date. In addition, it most likely would not be ethical. Estimates of the incidence of psittacosis are available, but this is not indicative of the asymptomatic presence of Cps DNA on the ocular adnexal tissue. A rough estimate of the annual incidence of psittacosis is 1 in 17 million people in the U.S., 1 in 3 million in the U.K., 1 in 3 million in France, 1 in 3 million in Italy, 0 in 16 million in the Netherlands, 1 in 4 million in Germany, 1 in 76 million in China, 1 in 7 million in Japan, 1 in 1 million in North Korea, 1 in 2 million in South Korea, and 1 in 1 million in Taiwan.24 These numbers are based on extrapolated statistics from the U.S., U.K., Canada, and Australia and do not take into account any genetic, cultural, or environmental differences; therefore, these statistics may be highly inaccurate and provide only general indications as to the actual total incidence of psittacosis.
Yet another possible explanation for the variability in Cps DNA prevalence noted across the studies may be that in all the studies reported to date, a potpourri of OAL histologic subtypes have been lumped together and the results are not stratified based on the histologic classification of the tumors. For example, it is possible that the prevalence of Cps DNA for conjunctival MALT lymphoma would be quite different than that for Mantle cell lymphoma or diffuse large cell lymphoma of the orbit or lacrimal gland.
The efficacy of antibiotics for the treatment of OAL was also found to be quite variable across the 4 reports published to date. The 2 consecutive reports by Ferreri et al.18, 19 contained what to our knowledge are the only 3 well‒documented cases of an objective response to antibiotics reported to date. These patients had received prior treatment with interferon, rituximab, radiation, or chlorambucil. The response in some of the patients in the study by Ferrari et al. was short‒lived, with lymphoma recurring in 6 of 7 patients during the first 12 months after treatment with antibiotics.19
To our knowledge, the only other study that reported 2 patients with OAL who achieved complete responses to antibiotic therapy was the study by Abramson et al.17 Unfortunately, no photographs or radiographic documentation of response was provided in this particular report. None of the other reports included radiographic or clinical evidence of response as documented by external or slit‒lamp photographs. In the study by Grunberger et al.,21 none of the 11 patients with OAL responded to antibiotic treatment with a follow‒up of > 9 months for all patients; all 11 patients had achieved a complete response after treatment with radiation or chemotherapy.
Judging the “response” of OAL can be problematic, and descriptions of response can be inaccurate unless documentation is provided in the form of imaging studies or photographs. Some patients with ocular adnexal MALT lymphoma may have had stable disease during antibiotic therapy simply because they had disease that was indolent enough that the period of observation was not long enough to permit documentation of any measurable progression. In other cases, particularly in cases of conjunctival MALT lymphoma, the disease load (lymphoma load) might have been very small or immeasurable because the bulk of the involved conjunctiva may have been removed during the diagnostic biopsy. Furthermore, the indolent nature and prolonged time to progression of MALT lymphoma of the conjunctiva may make the presence of “stable” disease in response to antibiotic treatment meaningless. Some authorities have even made the controversial recommendation to offer no treatment for the ocular component of this disease, given its indolent nature and prolonged time to disease progression.25 It is also possible and quite likely that the response rate for conjunctival MALT lymphoma to any treatment modality may be quite different from that for other forms of ocular adnexal lymphoma such as follicular, diffuse large cell, or mantle cell lymphoma; therefore, grouping these lymphomas together in studies that assess the efficacy of antibiotics is not a good strategy. A limitation of all studies of the efficacy of antibiotics for OAL published to date is that there is no stratification of data based on the histologic subtype of OAL. The response rate to any treatment and progression-free survival may be quite different for an indolent type of OAL such as conjunctival MALT lymphoma compared with a more aggressive histologic type such as Mantle cell lymphoma of diffuse large cell lymphoma.
In summary, our 2 meta‒analyses suggest a striking variability in the association between Cps and OAL across geographic regions and even among different studies from the same geographic regions and among studies published by the same group of investigators over time. The findings of the current study also suggest that antibiotics have relatively low and variable efficacy in the treatment of OAL. The variation in response to antibiotics, the lack of objective methods of assessment of response in the majority of reports published to date, the lack of stratification of response rates based on histologic subtypes of OAL, and the short follow‒up times suggest an unproven role for antibiotics in the treatment of OAL. Prospective studies with standard objective and measurable response criteria, a longer follow‒up time, and the stratification of patients based on their histologic subtype of OAL are needed to further clarify the role of antibiotics in the treatment of OAL.
- 2Marginal zone lymphoma. J Natl Compr Canc Netw. 2006; 4: 311–318., .
- 10Ocular adnexal lymphoma is highly associated with Chlamydia psittaci. Eur J Cancer. 2005;( suppl 3): 282–283. Abstract 982., , , et al.
- 24Adviware. Wrong diagnosis. http://www.wrongdiagnosis.com/p/psittacosis/stats-country.htm. Accessed June 18, 2007.