The Serotonin Transporter Gene Polymorphism (SLC6A4) and Risk for Psychiatric Morbidity and Comorbidity in the Baltimore ECA Follow-Up Study

Abstract

Background:

The human serotonin transporter (SERT) gene polymorphism (5HTTLPR) has been associated with multiple psychiatric disorders, including major depression, anxiety disorders, and substance use disorders. This study investigated the association between 5HTTLPR and psychiatric morbidity and comorbidity in a psychiatrist-examined population sample.

Methods:

628 participants, mean age 48.3 years old, were assessed by psychiatrists using the Schedules for Clinical Assessment in Neuropsychiatry. Associations between 5HTTLPR and the prevalence, comorbidity, and time-to-diagnoses for 16 psychiatric conditions were evaluated, using several analytical approaches.

Results:

The SERT S allele was significantly associated with an increased lifetime prevalence of panic disorder. There was a “protective” association between SERT gene S allele carrier status and the risk of obsessive-compulsive disorder (OCD) in time-to-event analysis. Carriers of the S allele had a significant increased risk of two specific comorbid disorder pairs: major depressive disorder (MDD) and social phobia, and MDD and agoraphobia. Overall, there was no increased risk of receiving an initial or an additional diagnosis for a mental disorder in the SERT S allele carriers Conclusions: The findings suggest that the S allele carrier status is associated with an increased prevalence of panic disorder in a community sample. There was an increased risk for comorbidity in a more homogeneous subgroup of cases with MDD and social phobia, as well as or agoraphobia. Our findings suggest a specific effect of the SERT promoter gene polymorphism on a subgroup of individuals identifiable by their comorbidity.

INTRODUCTION

The human serotonin transporter (SERT) is encoded by a single gene (SLC6A4) located on the long arm of chromosome 17 (17q11-17q12). The most investigated region of the SLC6A4 gene is the serotonin transporter gene promoter region (5-HTTLPR), located 1kb upstream of the SERT gene transcription initiation site. 5-HTTLPR has been the object of great attention due to a functional polymorphism comprised by a 44- base pair deletion /insertion in the 5’ regulatory region. The variant with the 44- base pair insertion is referred to as the long (L) allele, while the variant without the 44- base pair is called the short (S) allele [1-3]. The S variant of the 5-HTTLPR has been shown to have lower levels of gene transcription for SERT [4, 5].

Due to its impact in transcription, the 5-HTTLPR has been the object of multiple investigations that aimed at establishing its role as a risk factor for several psychiatric disorders. Multiple studies reported a significant association between the SERT S allele and an increased risk for depression, both as a candidate gene, or interacting with stressful events. [1, 6-18]. In contrast, several studies have failed to replicate this finding [19 -22]. In addition, a recent large study [23] conducted gene-wise association analyses for several previously studied candidate gene polymorphisms, and reported no association between the 5-HTTLPR and an increased risk for lifetime prevalence of major depressive disorder, or for an increased in the symptom severity of the depressive episode.

The SERT S carrier status also has been associated with an increased risk for eating disorders [24] [23], post-traumatic stress disorder [25], and alcohol dependence [26]. In contrast, an increased risk for obsessive-compulsive disorder was has been associated with the SERT L allele carrier status [27, 28].

In summary, there is a vast body of literature on the potential association of 5-HTTLPR with an extensive list of psychiatric conditions. The results of these investigations have been, so far, inconclusive. Considering the role of serotonin transmission in the central nervous system, and the role of the SERT in the regulation of serotonin transmission, the existence of a functional polymorphism in the gene that codes for serotonin transporter continues to be an area of active interest.

The aim of this investigation was to study the association between the 5-HTTLPR with the lifetime prevalence, the comorbidity burden, and the time to diagnoses of sixteen mental disorders. In addition, we examined the association of 5-HTTLPR with the overall risk of developing psychopathology in a community-based sample, and assessed by board certified or eligible psychiatrists using a standardized diagnostic instrument. Using this comprehensive approach, we avoided one of the limitations of candidate genes and their association with a single categorical mental disorder. To pursue these aims, we used data collected during the Epidemiologic Catchment Area Follow-up Study in Baltimore [11, 29].

2. METHODS:

2.1. PARTICIPANTS

The participants for this study were sampled from the Baltimore Epidemiologic Catchment Area Follow-up Study. The Baltimore cohort of the Epidemiologic Catchment Area (ECA), a population-based study that assessed the burden of psychopathology in the community. The target population for the Baltimore ECA site consisted of the household residents of East Baltimore (an area with 175,211 adult inhabitants in 1981). Out of the 4238 individuals that were probabilistically designated for interview in the first wave, 3,481 participants (82% of 4,238 designated in the ECA sampling process) were interviewed with the Diagnostic Interview Schedule (DIS) [30] at the Baltimore site [11, 29]]. In 1982, 2,768 subjects (73% of the first wave) of the ECA original cohort were re-interviewed for the second wave. This cohort was followed in two additional waves (Wave 3=1993–1996 and Wave 4=2003–2004) as part of the Baltimore ECA Follow-up Study (EFU) [19,11,29]. The Baltimore ECA cohort was the source of several investigations including the Clinical Reappraisal (CR) study [31], and the Hopkins Epidemiology of Personality Disorder Study (HEPS) [32-34]. For the current study, a subsample of participants of the EFU were selected by a process that has been previously described [33,34]. Briefly, 442 ECA participants who had a lifetime diagnosis of mania, depression, panic disorder, obsessive-compulsive disorder, alcohol use disorder, or drug use disorder in previous ECA waves or the CR were included. The sample was complemented with an addition of 186 randomly selected ECA participants. This resulted in 628 subjects who had completed a psychiatric interview using Schedules for Clinical Assessment in Neuropsychiatry (SCAN) [35, 36] and who had provided a sample for genetic testing. Age of participants ranged from 30 to 90 years old (Mean age was 48.3 years old). The study was approved by the Institutional Review Board of Johns Hopkins Medical Institutions. Prior to the interview, participants provided written informed consent for study procedures, including the interview and collection of a DNA sample.

2.2. MEASURES

2.2.1. PSYCHIATRIC INTERVIEW

Board-certified or board-eligible psychiatrists assessed participants using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) [35, 36]. All psychiatrists had undergone a SCAN training course at an official World Health Organization SCAN training center. For the EFU, several items were added to the SCAN, version 1.0. Some of these additions were incorporated into the second edition of the SCAN [36]. The interviews were conducted between 1993 and 1999. Participants were diagnosed by trained psychiatrists using DSM-III-R criteria [37].

2.2.2. GENETIC SAMPLE

Participants provided either a venous blood or a buccal mucosa sample. Blood samples collected were placed onto a specially formulated “Isocode” Card. DNA was isolated from peripheral blood leukocytes using Puregene Blood Kit chemistry on an Autopure LS automated DNA purification instrument (Qiagen, Valencia, CA). Following manufacturer’s protocols, blood collected on Isocode Cards was isolated by heating hole punches in distilled water at 95 degrees centigrade for 30min. Buccal cell swabs were manually isolated using Puregene DNA isolation kit (Qiagen). Manufacturer’s protocols were used in this procedure. DNA concentrations were established by spectrophotometry using a DU 530 Life Science UV/Vis Spectrophotometer (Beckman Coulter, Brea, California) [34].

The serotonin transporter genotype was determined by polymerase chain reaction amplification [34]. For the analysis, we combined the “S/S” and “L/S” genotypes into a carrier group and compared this group to the non-carrier group that only included those with the “LL” genotype. This grouping was based on previous literature that had analyzed the functional effect of the S carrier allele, both in vitro and in vivo [34].

2.3. ANALYTICAL APPROACH

2.3.1. LIFETIME PREVALENCE OF MENTAL DISORDERS

Using Poisson regression analysis, we compared carriers of the S allele of the SERT polymorphism with non-carriers to determine whether S allele carriers had an increased lifetime prevalence of the most frequently diagnosed mental disorders in the study sample. These sixteen disorders included alcohol dependence, alcohol abuse, opioid dependence, schizophrenia and other psychotic disorders, major depressive disorder, bipolar disorder, panic disorder, agoraphobia, social phobia, obsessive-compulsive disorder, generalized anxiety disorder, simple phobia, dysthymic disorder, anxiety disorder not otherwise specified, depressive disorder not otherwise specified, and adjustment disorder with depressed mood.

2.3.2. COMORBIDITY OF MENTAL DISORDERS

Using logistic regression analysis, we estimated the odds ratio (OR) of the comorbidity associations between the sixteen mental disorders in SERT S allele carriers versus non-carriers. For those affected by a selected index condition, the calculated ORs obtained represent the ratio between of the odds of being diagnosed with a selected second condition for S allele carriers and the odds of being diagnosed with that second condition for non-carriers.

2.3.3. OVERALL RISK FOR PSYCHOPATHOLOGY

Using three complementary approaches, we analyzed the association between the 5-HTTLPR and the overall risk for increased psychopathology. The overall psychopathology risk was defined as the presence of one or more diagnoses of any of the 16 mental disorders in the lifetime of the participants. In the first approach, using chi square and Fisher’s exact tests, we examined the association of the SERT S allele carrier status and the number of lifetime prevalence diagnoses received.

In the second approach, we analyzed the association of the SERT S allele carrier status and the risk of receiving a diagnosis for any mental disorder. We performed logistic regression to obtain an odds ratio to determine the risk of this association. We adjusted the regression models for the effect of sex, age, marital status, and ethnicity, as these are the most relevant confounding factors cited in the literature. In the third approach, we analyzed the association of the SERT S allele carrier status and the risk of receiving a second diagnosis for any mental disorder in those participants who already had received one diagnosis. We performed a logistic regression analysis to obtain an odds ratio to determine the risk for this association. We adjusted the regression models for the effect of potential confounding variables (sex, age, marital status, and ethnicity).

2.3.4. TIME TO DIAGNOSIS

Using the Kaplan-Meier approach, we calculated the events (lifetime diagnosis) and exposure time (number of years until being diagnosed) both for the carriers of the SERT S allele and for the non-carrier participants. For each disorder, we obtained separate incidence rates for carrier and non-carrier groups, and the ratios between these incidence rates. The ratios were a measure of the association of the SERT S allele carrier state with the time to onset of the examined mental disorder. Thus, this analytical approach enabled us to examine whether the SERT S allele carrier status was associated with receiving a diagnosis for a mental disorder at an earlier age.

Log-Rank tests for each condition were calculated to estimate whether the time to diagnosis curves (survival curves) differed between groups: namely, whether the carriers had a different time trajectory in attaining each diagnosis under examination.

Finally, we performed Cox proportional hazards regression to examine the association between the SERT S allele carrier status and the time, in years, to attain diagnosis. The Cox proportional hazards model also allowed adjustment of hazard estimates by potential confounding variables: namely, sex, ethnicity, and marital status. Thus, using this technique we were able to obtain a hazard ratio between the carrier and non-carrier participants for each condition examined, both unadjusted and adjusted for the aforementioned variables. Lastly, we tested whether our analysis for each disorder met the proportional-hazards assumptions required by the Cox regression model.

3. RESULTS:

3.1. GENERAL CHARACTERISTICS

A total of 628 participants provided a DNA sample that was genotyped for the 5-HTTLPR. Participant’s mean age was 48.3 years old (95% CI: 47.4; 49.2), 64% were Female, 42% were married. The frequencies of 5-HTTLPR genotypes in the sample were “L/L”, 280 (44.6%); “L/S”, 274 (43.6%); and “S/S”, 74 (11.8%). The distribution of the alleles followed the Hardy-Weinberg Equilibrium chi² (1df): 0.38 (p>0.5).

3.2. LIFETIME PREVALENCE OF MENTAL DISORDERS

We estimated the lifetime prevalence (LP) of alcohol dependence, alcohol abuse, opioid dependence, schizophrenia and other psychotic disorders, major depressive disorders (MDD), bipolar disorder, panic disorder, agoraphobia, social phobia, obsessive compulsive disorder (OCD) , generalized anxiety disorder (GAD), simple phobia, dysthymic disorder, anxiety disorder not otherwise specified, depressive disorder not otherwise specified, and adjustment disorder with depressed mood in the sample using Poisson regression models and chi square tests.

Table 01 summarizes the un-weighted LP estimates with 95% confidence intervals (CI) of the 16 above- mentioned conditions. The disorders with the highest unweighted prevalence (%) in the entire sample were: alcohol dependence (20.2), followed by MDD (18.8), simple phobia (14.2), and social phobia (11.6).. Table 01 displays the prevalence ratio estimates between the carrier and the non-carrier subsamples with 95% Confidence Intervals. Panic disorder was the only condition examined in the study sample that had a significant prevalence ratio (2.61 (95% CI: 1.20; 5.69), p=0.02). Carriers of the S allele SERT gene polymorphism had risk for panic disorder in their lifetime that was 2.61 times higher than those who were non-carriers.

Table 01.

Lifetime Prevalence Estimates as Percentages, for the Whole Sample, and by SERT (S) Carrier Status

Condition	Whole Sample (n=628)	SERT Carrier (SS+SL) (n=348) (54.4%)	SERT Non-Carrier (LL) (n=280) (44.6%)	Carrier/ Non-Carrier Ratio
Alcohol Dependence	20.2	21	19.3	1.1(0.8-1.5)
Alcohol Abuse	6.4	6.6	6.1	1.1(0.6-2)
Major Depressive Disorder	18.8	18.4	19.3	1(0.7-1.3)
Bipolar Disorder	2.4	3.2	1.4	2.2(0.7-6.9)
Psychotic Disorders	1	0.9	1.1	0.8(0.2-4)
Panic Disorder	5.4	7.5	2.9	2.6(1.2-5.7) *
Agoraphobia	5.9	5.2	6.8	0.8(0.4-1.4)
Social Phobia	11.6	12.6	10.4	1.2(0.8-1.9)
Obs Comp DO	1.3	0.6	2.1	0.3(0.1-1.3)
Gen Anx DO	2.9	2.3	3.6	0.6(0.3-1.6)
Anxiety Disorder NOS	2.1	2	2.1	0.9(0.3-2.8)
Dysthymic Disorder	1.3	1.2	1.4	0.8(0.2-3.2)
Depressive DO NOS	2.2	2	2.5	0.8(0.3-2.3)
Simple Phobia	14.2	12.6	16.1	0.8(0.5-1.2)
Adj DO Depressed Mood	4	4	3.9	1(0.5-2.2)
Opioid Dependence	4	4	3.9	1(0.5-2.2)

Table 01 displays the un-weighted LP prevalence estimates as percentages with 95% confidence intervals (CI) of the 16 mental and substance use disorders examined. Column 01 displays the un-weighted LP estimates for the whole sample (N=628). Column 02 displays the un-weighted LP estimates for the SERT gene (s) allele carrier sub-sample (N=348). Column 03 displays the un-weighted LP estimates for the SERT (s) allele non-carrier sub-sample (N=280). Column 04 displays the prevalence ratio estimates between the carrier and the non-carrier subsamples with 95% Confidence Interval (95% CI).

^*Marks Ratios that achieved statistical significance (p<0.05)

3.3. COMORBIDITY BETWEEN MENTAL DISORDERS

The odds ratios (ORs) of the SERT S carrier status and the risk of comorbidity are displayed in Table 02. In several cases, the analyses could not be performed, due to insufficient numbers of affected participants. The carrier status was associated with a significant increase in the odds for comorbidity in three comorbidity pairs.

Table 02.

Association of SERT (s) Carrier Status with Comorbidities, Whole Sample (N=628)

Index Condition	Alcohol Dependence	Alcohol Abuse	Major Depressive DO	Bipolar Disorder	Psychotic Disorders	Panic Disorder	Agoraphobia	Social Phobia	Obsessive Comp DO	Gen Anxiety DO	Anxiety DO NOS	Dysthymic Disorder	Depr DO NOS	Simple Phobia	Adjust DO Dep Mood	Opioid Dependence
Alcohol Dependence	N/A	1.5	0.6	3.1			1.6	1.2		0.2	0.7		0.2	0.5	2.3	1.3
Alcohol Abuse	1.6	N/A	0.4			3.4		1.6						0.2		0.7
Major Depressive Disorder	0.7	0.5	N/A			2.0	2.3	5.1*	0.3	0.7	0.4	1.7		1.0		1.4
Bipolar Disorder	1.7			N/A			0.1	0.2
Psychotic Disorders					N/A			1.0
Panic Disorder		1.3	0.4			N/A	0.7	3.1	0.3						0.3	0.6
Agoraphobia	3.5		4.7*	0.5		3.3	N/A	1.1						0.6		1.1
Social Phobia	1.0	1.4	4.8*	0.6	0.7	6.2	0.6	N/A	0.2	1.4			0.7	0.7	1.3	2.1
Obsessive Compulsive DO			1.0			5.0		0.5	N/A
Generalized Anxiety Disorder	0.2		1.0					4.0		N/A				1.4
Anxiety Disorder NOS	0.8		0.3								N/A			2.0
Dysthymic Disorder			3.0									N/A		1.0
Depressive Disorder NOS	0.2							1.0					N/A			2.4
Simple Phobia Disorder	0.7	0.2	1.3				0.7	1.3		1.0	2.1	1.0		N/A	1.0	2.1
Adjust DO Depressed Mood	2.7					0.8		1.7						0.7	N/A
Opioid Dependence	1.5	0.8	1.5			1.7	0.8	2.7					1.7	1.7		N/A

Table 02 summarizes the ratio of the odds of the SERT (S) allele carriers to the odds of the non-carriers, of the comorbidity for the conditions listed in the column headings (Condition B), on those subjects who have been diagnosed with the index condition listed in the row headings (Condition A or Index). Blank cells represent associations with insufficient numbers to complete calculation.

^*Marks Ratios that achieved statistical significance (p<0.05)

For subjects diagnosed with MDD, the carriers of the S allele, had 5.1 times the risk for being diagnosed with social phobia as compared to the risk of non-carriers (OR: 5.1 (95%CI: 1.8; 14.7)). For subjects diagnosed with agoraphobia, the carriers of the S allele, had 4.7 times the risk for being diagnosed with MDD as compared to the risk of non-carriers (OR: 4.7 (95%CI: 1.1; 19.8)). Finally, for subjects diagnosed with social phobia, the carriers of the S allele, had 4.8 times the risk for being diagnosed with MDD as compared to the risk of non-carriers (OR = 4.8 (95%CI: 1.6; 14.8)).

It is important to highlight that, as we considered the value of the comorbidity OR in relation to a third variable (the 5-HTTLPR), the OR values varied according to which condition was considered the first or index condition. In table 03, the value of the association MDD (A) ->social phobia (B), differs from the one of social phobia (A) -> MDD (B). This is explained by the fact that in the first analysis we are comparing the genetic information of the comorbid cases of MDD and social phobia with the genetic information of the non-comorbid cases of MDD, whereas, in the second analysis we are comparing it with the genetic data of the non-comorbid cases of social phobia.

Table 03.

Odds Ratio Comorbidity for the Significant Comorbid Associations

Grouping	SERT Carrier (ss+sl) Individuals Affected (%)	SERT Non-Carrier (II) Individuals Affected (%)	Total Number Individuals Affected
Association Major Depressive Disorder (A) → Social Phobia (B)
Comorbid Cases MDD-Social Phobia	22 (81.5)	5 (18.5)	27
Non- Comorbid Cases MDD	42 (46.2)	49 (53.8)	91
Totals	64 (54.2)	54 (45.8)	118
Odds Ratio = 5.1 (95%CI: 1.8; 14.7)
Association Social Phobia (A) => Major Depressive Disorder (B)
Comorbid Cases Social Phobia- MDD	22 (81.5)	5 (18.5)	27
Non- Comorbid Cases Social Phobia	22 (47.8)	24 (52.2)	46
Totals	44 (60.3)	29 (39.7)	73
	Odds Ratio = 4.8 (95%CI: 1.6; 14.8)
Association Agoraphobia (A) => Major Depressive Disorder (B)
Comorbid Cases Agoraphobia- MDD	10 (71.4)	4 (28.6)	14
Non- Comorbid Cases Agoraphobia	8 (34.8)	15 (65.2)	23
Totals	18 (48.6)	19 (51.4)	37
Odds Ratio = 4.8 (95%CI: 1.6; 14.8)

Table 03 displays the two by two table and analysis for the Odds Ratio for the significant comorbid associations between SERT (s) allele carriers and non-carriers

3.4. NUMBER OF LIFETIME DIAGNOSES

Table 04 displays the results of the association between the S allele carrier status and the overall number of lifetime diagnoses. There were no significant differences in the distribution of the number of lifetime diagnoses between the (s) allele carriers and the non-carriers (Chi Square (6) p: 0.93 /Fisher's Exact Test p: 0.97)

Table 04.

Association of the SERT Gene Polymorphism with the Number of Lifetime Diagnoses

Grouping	SERT Carrier (ss+sl) Individuals Affected (%)	SERT Non-Carrier (II) Individuals Affected (%)	Total Number Individuals Affected
Association of SERT Carrier Status with Prevalence of One or More Diagnoses LP Diagnosis (N=628)
No Diagnosis LP	120 (44.9)	147 (55.1)	267
One or More Diagnosis LP	160 (44.3)	201 (55.7)	361
Totals	280 (44.6)	348 (55.4)	628
OR Unadjusted: 1.0 (0.85 - 1.4) p: 0.88 / OR Adjusted: 1.0 (0.7 - 1.4) p: 0.92
Association of SERT Carrier Status with Prevalence of Two or More Diagnoses LP Diagnosis (N=361)
One Diagnosis LP	82 (45.1)	100 (54.9)	182
Two or More Diagnosis LP	78 (43.6)	101 (56.4)	179
Totals	160 (44.3)	201 (55.7)	361
OR Unadjusted: 1.1 (0.7-1.6) p: 0.78 / OR Adjusted: 1.0 (0.7-1.6) p: 0.84

Table 04 displays the association of the SERT Gene Polymorphism with the Number of Lifetime Diagnoses. ORs were adjusted for gender, age, ethnicity, and marital status.

No significant differences were found when participants were grouped into dichotomous categories. The SERT S allele carrier status was not associated with an increase in the odds of receiving an initial diagnosis as compared to non-carriers (OR Unadjusted: 1.0 (0.8 - 1.4) p: 0.88 / OR Adjusted: 1.0 (0.7 - 1.4) p: 0.92). In addition, the SERT S allele carrier status was not associated with an increase in the odds of receiving a second diagnosis in participants who already had received an initial diagnosis, as compared to non-carriers (OR Unadjusted: 1.1 (0.7-1.6) p: 0.78 / OR Adjusted: 1.0 (0.7 - 1.6) p: 0.84)

In addition, in a separate sensitivity analysis, we excluded all NOS disorders and adjustment disorders from the analysis because these conditions do not have a clear diagnostic definition. We also excluded psychotic, bipolar, and drug use disorders from the analysis. The rationale for this sensitivity analysis was to examine the association of SERT gene polymorphism and the risk of depressive and anxiety disorders, as this was the main focus of most previous literature on this allele. Thus, we examined the association of the SERT S allele carrier status with the risk of receiving an initial or an additional mood or anxiety disorder diagnosis. These analyses yielded non-statistically significant results.

3.5. TIME TO DIAGNOSIS (SURVIVAL ANALYSIS)

Table 05 summarizes the estimates of several analyses performed comparing the SERT S allele carrier with the non-carrier subgroups (incidence rates ratios, long rank tests, as well as the unadjusted and adjusted Cox proportional hazards regression analysis results). The SERT S allele carrier status was associated with 2.7 times the risk for receiving a diagnosis of panic disorder (incidence rate ratio: 2.7 (95% CI: 1.1-7.4). For panic disorder subjects, this difference between carriers and non-carriers was also observed in the log-rank test, and in the unadjusted hazard ratio. The hazard ratio, adjusted for sex, ethnicity, and marital status, was of borderline significance (adjusted hazard ratio: 2.4 (95% CI: 1.0-2.5) (Table 05).

Table 05.

Time to Event Estimates SERT S Allele Carriers vs Non-Carriers (Non-Carriers N=266/ Carriers N=336)

	Incidence Rate Ratio (IRR) (95% CI)	Log Rank P value	Unadjusted Hazard Ratio (95%CI)	Adjusted Hazard Ratio (95%CI)
Alcohol Dependence	1.1 (0.8-1.5)	0.72	1.1 (0.8-1.5)	1.1 (0.75-1.48)
Major Depr DO	0.9 (0.6-1.4)	0.76	0.9 (0.7-1.4)	0.9 (0.7-1.4)
Bipolar DO	1.0(0.2-49)	0.98	1.0 (0.3-3.7)	0.8 (0.2-3.1)
Psychotic Dos	0.8(0.1-10.7)	0.82	0.8(0.15.6)	0.9(0.16.5)
Panic Disorder	2.7 (1.1-7.4)*	0.02*	2.6 (11-6.1)*	2.4(10-5.5)*
Agoraphobia	0.8(0.4-16)	0.45	0.8(0.4-15)	0.8(0.4-15)
Social Phobia	13(0.8-2.2)	0.21	14(0.8-2.3)	14(0.8-2.3)
Obs Comp DO	0.1 (0.0-11)*	0.03*	0.1 (0.2-11)*	0.1 (0.0-10)*
Gen Anxiety DO	0.8 (0.3-2.4)	0.63	0.8 (0.3-2.1)	0.8 (0.3-2.2)
Simple Phobia	0.7(0.5-12)	0.19	0.7 (0.5-12)	0.9(0.5-14)
Opioid Dependence	0.6 (0.1-3.4)	0.49	0.6 (0.1-2.6)	1.1 (0.2-4.7)

Table 05 summarizes the estimates of several survival analyses techniques comparing the SERT (s) allele carrier with the non-carrier subgroups. Column 01 included the estimated the Ratio between the Incidence Rates of the two groups. Column 03 includes the results of the Log Rank tests comparing the survival curves of the two groups. Column 04 displays the results of the unadjusted Hazard Ratios for each disorder obtained from the Cox Proportional Hazards Regression Analyses. Column 06 summarizes the estimates of the Hazard Ratios, adjusted by gender, ethnicity, and marital status.

^*Marks Ratios that achieved statistical significance (p<0.05)

In addition, the SERT S allele carrier status was associated with a significant 87% decrease in risk for receiving a diagnosis of OCD (incidence rate ratio: 0.1 (95%CI: 0.0-1.0). For OCD, this difference between carriers and non-carriers was also significant using the log-rank test, and both the unadjusted hazard ratio (HRR), as well as for the HRR adjusted for gender, ethnicity, and marital status (Adjusted HRR: 0.1 (95% CI: 0.0-1.0) (Table 05).

4. DISCUSSION

We explored the association of the 5-HTPPLR and the risk for psychopathology using several complementary analytical approaches. We examined the association of 5-HTPPLR with the lifetime prevalence, the comorbidity burden, and the time to diagnoses for 16 mental disorders. In addition, we examined the odds of receiving an initial or an additional mental disorder diagnosis in the SERT S allele carriers.

Our findings suggest a significant association between SERT S allele carrier status and an increased lifetime prevalence of panic disorder. Previous literature focusing on the association between the 5-HTTLPR and panic disorder yielded negative results. Strug et al. [38] reported no association of the 5-HTLLPR with an increased risk for panic disorder in a sample of N=179 subjects sampled from the NIMH Human Genetic Initiative. In addition, Blaya et al. [39] reported no association of the 5-HTLLPR with panic disorder in a case-control candidate gene association study with 107 patients. Potential explanations for our findings could be the differences in the source of our study participants (community vs. clinic-based), as well as the limited sample size of these studies. Our findings are consistent with the serotonin deficit theory proposes that serotonin has a restraining effect on panic behavior, and a 5-HT deficit may facilitate panic [40].

As described in the introduction section, several studies have examined the association of 5-HTTLPR with depression, both as a candidate gene, and in a gene by-environment interaction hypothesis. The results have been inconsistent. This study did not find an association between 5-HTTLPR and an increase in the lifetime prevalence of major depressive disorder (MDD). However, we found a significant association between the S allele carrier state and an increased risk for comorbidity for MDD and social phobia, as well as for MDD and agoraphobia. We believe that these findings could be explained by the fact that these three conditions (as well as their respective co-occurrences) were highly prevalent. Thus, there was a critical number of subjects in the sample that allowed for any heterogeneity between SERT carriers and non-carriers to manifest itself in a significant way. The finding in a specific group of persons within the MDD category, but not all persons with major depression could explain the inconsistent findings described above. It also emphasizes the importance of comprehensive phenotyping in research participants.

We did not find an association between the SERT S allele carrier status and the odds of receiving an initial, or an additional diagnosis for of any mental disorder This finding suggests that there is not an association of 5-HTTLPR with an increased risk for psychopathology, when we consider the cumulative or overall number of diagnoses. Thus, we should characterize any risk conveyed by the 5-HTTLPR in developing psychopathology as specific and modest rather than broad and robust.

The time to diagnoses analytical approach supported the lifetime prevalence findings. The SERT S allele carrier status was associated with an increased risk for of receiving a diagnosis of panic disorder. In addition, the SERT S allele carrier status was associated with a significant decreased risk for receiving a diagnosis of OCD. These findings are consistent with previous literature. The 5-HTLLPR has been associated with OCD in case controls and family-based studies. In a case control study that included 75 cases of Caucasian OCD patients, Bengel et al. [28] reported that the OCD subjects were more likely to carry two copies of the long SERT L as compared to the 397 matching controls. Bloch et al. (2008) [27] reported a significant association between the L allele and OCD. Sinopoli et al [41] reported a significant association between the [LG + S] variant and hoarding in males with OCD HTTLPR (the L allele modelled in a bi-allelic fashion as LA and LG). In a recent meta-analysis, Taylor et al. (42) reported a significant association between the 5-HTTLPR and an increased risk for OCD. In addition, the La allele was also associated with an increased risk for the comorbidity between OCD and depression, posttraumatic stress disorder, and substance use disorder. In a recent study

Contemporary approaches in psychopathology posit the existence of a meta-structure that could explain the substantial degree of observed comorbidity between major mental disorders [43, 44, 45]. Using several analytical approaches, we have found associations between the 5-HTTLPR and panic disorder, MDD, social phobia, agoraphobia, and OCD. Thus, our findings suggest that the 5-HTTLPR could be linked to modest increase in the vulnerability to conditions that have been categorized into a high-order grouping of internalizing disorders [43, 44, 45].

Strengths and Limitations

The study cohort was the result of a careful sampling methodology, it was representative of the Baltimore metropolitan region. Participants were comprehensively assessed for numerous mental disorders by trained psychiatrists using the SCAN. In addition, the genetic information on the participants enabled us to examine the association of the 5-HTLLPR with multiple mental disorders. The sample size might appear modest at first, especially when examining conditions with low prevalence. Nevertheless, it provides an important perspective on the prevalence of several mental disorders in the community. We were able to analyze information of the morbidity of multiple mental disorders simultaneously in the same cohort. Thus, we were able to study the phenomenon of comorbidity in a sample without the biases that we might find in treatment clinical samples.

Additional research has challenged the concept of a bi-allelic SERT polymorphism [43], reporting ten sequence variants of the 5-HTLLPR [40]. In addition, a third polymorphism in the SLC6A4 gene, consisting of a variable number of tandem repeat (VNTR) in intron 2 (STin2 VNTR) has been reported [41-46]. In our study, we considered the SERT promoter region polymorphism in a bi-allelic approach. We were unable to explore a more novel tri-allelic approach for the 5-HTTLPR, as well as additional single nucleotide polymorphisms present in the SERT gene.

Due to the limited size of our sample, we were unable to analyze the association of the 5-HTTLPR with several mental disorders controlling for traumatic events. Thus, we were unable to test gene-by-environment interaction [GxE] effects of this polymorphism.

Our sample was obtained from only one geographical region, and has a multi-ethnic composition. This could have introduced population stratification bias in our findings.

In addition, the estimation of the association between the 5-HTTLPR and several mental disorders using multiple analytical approaches required several tests. Thus, we should consider the potential error resulted from multiple testing. The significance testing for single comorbid associations should be considered an exploratory endeavor. Future studies including larger samples should examine the specific comorbid pairs that were found to have a positive association with the SERT S carrier status to confirm our findings.

Candidate gene studies are in disfavor for several reasons. This is emphasized in the paper by Border at al. [23]. They conducted a GWAS study in a large sample examining several reputed candidate genes without identifying any association, concluding that candidate studies are flawed. However, several factors need be considered. First, diagnostic heterogeneity and phenotype assessment may influence the findings. Our study takes these issues into account. Second, association in GWAS studies of the 5-HTTLPR polymorphism may be missed as the 5-HTTLPR variant is not a SNP and therefore not directly assayed in genome-wide arrays [46]. Furthermore, the functional impact of this polymorphism remains an important aspect in understanding psychopathology.

5. CONCLUSIONS

Our findings suggest an increased LP of panic disorder in participants who were carriers of the SERT gene S allele. There was a “protective” association between SERT gene S allele carrier status and the LP of OCD in the time-to-diagnosis or survival analysis. Our results indicate a specific effect of the 5-HTTLPR on a subgroup of patients identifiable by their comorbidity. The carriers of the S allele had an increased risk for two comorbidity pairs, MDD and social phobia, and agoraphobia and MDD. Considering the overall risk for psychopathology, there was no increased risk of the SERT S allele carriers in receiving an initial, or an additional diagnosis for any mental disorder

Future work should include all possible functional variants of the 5-HTTLPR.and focus on large population- based cohorts. Research should emphasize assessment for a wide number of conditions, controlling for environmental interactions and using standardized instruments. Comorbidity between depressive and anxiety disorders is widely prevalent. Future work should focus on the possible role of 5-HTTLPR in an increased risk for comorbidity between these conditions

Considering the role of serotonin transmission in the central nervous system, a functional polymorphism in the gene that codes the SERT should continue to be an area of active research

Highlights.

• In a sample of the ECA study, we found an association of the SERT short allele was with an increased risk for panic disorder, and a decreased risk for OCD

• The SERT S allele was associated with an increased risk of two comorbidities: major depressive disorder (MDD) and social phobia, and MDD and agoraphobia.

• The SERT S allele was not associated with an increased risk for psychopathology

ABBREVIATIONS:

ECA

Epidemiologic Catchment Area Study

EFU

Baltimore Epidemiologic Catchment Area Follow-up Study

SERT

human serotonin transporter

5HTTLPR

serotonin transporter gene polymorphism

SCAN

Schedules for Clinical Assessment in Neuropsychiatry

LP

lifetime prevalence

MDD

major depressive disorder

OCD

obsessive compulsive disorder

OR

Odds Ratio