Given current evidence supporting a genetic predisposition for pelvic organ prolapse, we conducted a systematic review of published literature on the genetic epidemiology of pelvic organ prolapse. Inclusion criteria were linkage studies, candidate gene association and genome-wide association studies in adult women published in English and indexed in PubMed through Dec. 2012, with no limit on date of publication. Methodology adhered to the PRISMA guidelines. Data were systematically extracted by 2 reviewers and graded by the Venice criteria for studies of genetic associations. A metaanalysis was performed on all single nucleotide polymorphisms evaluated by 2 or more studies with similar methodology. The metaanalysis suggests that collagen type 3 alpha 1 (COL3A1) rs1800255 genotype AA is associated with pelvic organ prolapse (odds ratio, 4.79; 95% confidence interval, 1.91–11.98; P = .001) compared with the reference genotype GG in populations of Asian and Dutch women. There was little evidence of heterogeneity for rs1800255 ( P value for heterogeneity = .94; proportion of variance because of heterogeneity, I 2 = 0.00%). There was insufficient evidence to determine whether other single nucleotide polymorphisms evaluated by 2 or more papers were associated with pelvic organ prolapse. An association with pelvic organ prolapse was seen in individual studies for estrogen receptor alpha (ER-α) rs2228480 GA, COL3A1 exon 31, chromosome 9q21 (heterogeneity logarithm of the odds score 3.41) as well as 6 single nucleotide polymorphisms identified by a genome-wide association study. Overall, individual studies were of small sample size and often of poor quality. Future studies would benefit from more rigorous study design as outlined in the Venice recommendations.
Pelvic organ prolapse (POP) affects 40% of postmenopausal women and directly impacts bladder and bowel function, as well as quality of life. Surgical correction of POP is anticipated to increase 48% from 2010 to 2050 given the aging population in the United States. The pathophysiology of this prevalent disorder is believed to be multifactorial, involving vaginal parity and other obstetric risk factors, as well as advanced age, increased body mass index, smoking, constipation, and vaginal hysterectomy. Yet, even with multiple risk factors, there is a large component of risk that is not understood. This is exemplified by the fact that nulliparous women can develop prolapse, and conversely, most parous women do not develop prolapse. It is plausible that genetics contribute significantly to the development of prolapse. Studies show a 5-fold increased risk of prolapse among siblings of women with severe prolapse as compared with the general population and a high concordance of prolapse in twins, as well as, in nulliparous and parous sister pairs.
The interrelationship of epidemiologic, environmental, and genetic risk factors for POP constitutes the genetic epidemiology of prolapse. With improved understanding of these relationships, there may be a role for individual risk assessment in future. Perhaps, women at high-risk for prolapse may choose to prophylactically perform pelvic muscle strengthening exercises; or, following the development of prolapse, potentially opt for a primary sacrocolpopexy with mesh instead of pelvic reconstruction with native tissue. Currently, both our understanding of the genetic epidemiology of POP as well as our knowledge about the efficacy and longevity of treatment options is too limited to make definitive recommendations; but, as our knowledge advances, this information may be incorporated into patient counseling and treatment decisions. This type of personalized medicine is becoming a reality in other fields, such as cardiology and oncology, in which genetic risk stratification is more advanced. Given the preliminary data supporting a genetic component to the cause of prolapse, this study aimed to systematically review and highlight current research in this area. We focused on genome-wide association studies (GWAS), linkage, and candidate gene association studies in adult women with POP.
Methods for review
We initially conducted a broad search on the genetics and genetic epidemiology of POP and urinary incontinence. For the analysis presented here, only those papers pertaining to the genetic epidemiology of POP were included. Methodology adhered to the PRISMA statement guidelines. The search was limited to publications in English with an adult female population that were indexed in PubMed through Dec. 2012. There were no limitations on date of publication. Controlled vocabulary terms served as the foundation of our search with 1 clinical term (pelvic organ prolapse, cystocele, rectocele, urinary incontinence, urge incontinence, stress incontinence, mixed incontinence, pelvic floor, uterus/uterine/vaginal/vault, urogenital/bladder/pelvic organ/genitourinary, and prolapse, vaginal and defect, or enterocele) and 1 genetic term (genetic phenomena, genetics, genetic models, genetic techniques, polymorphism, genome, phenotype, genotype, gene, genes, variant, exome, exon, gene expression, microarray, sequencing, protein biosynthesis, protein, protein, proteomic, hereditary, familial or inherited). We excluded all newspaper articles, letters, comments, case reports, reviews, practice guidelines, news, historical articles, metaanalyses, legal cases, published erratum and congresses. References from key articles were hand-searched to identify additional studies.
We defined POP as anatomic prolapse of the vaginal walls and/or uterus and defined genetic epidemiology to include linkage studies, candidate gene association studies and GWAS. For GWAS and candidate gene studies, studies needed to include a comparator of women without prolapse. Outcomes of interest were single nucleotide polymorphisms (SNPs) associated with POP.
We used 4 reviewers: 2 MD clinicians and 2 PhD genetic epidemiologists. All reviewers evaluated the first 50 abstracts to ensure consistency. The remaining abstracts underwent dual review to determine inclusion or exclusion, followed by dual full-text review of all articles selected for inclusion. Discordance was resolved by third-party adjudication. Data from included articles were extracted using a standardized form and a second team member ensured the extraction was accurate, complete, and consistent. Given the translational nature of the project, dual review and data extraction involved a clinician and a genetic epidemiologist at each step in the process. Individual reviewers recused themselves from the evaluation and data extraction of any study they were involved in or had coauthorship. This study did not involve human subjects and was exempt from Institutional Board Review.
We graded the quality of the GWAS and candidate gene studies using the Venice guidelines. These guidelines grade the cumulative evidence in support of a genetic association based on 3 criteria: (1) the amount of evidence, (2) whether replication was performed, and (3) protection from bias. Each category can receive a grade of “A,” “B,” or “C”. Studies graded AAA have the strongest evidence, “A” and “B” studies indicate moderate evidence and any study with a category “C” represents only weak evidence.
A metaanalysis was performed of all SNPs evaluated by 2 or more studies with similar methodology. We used odds ratios (ORs) as the effect measure of choice to report the weighted associations between SNPs of interest and POP. If any study in a metaanalysis set reported only crude numbers and a χ 2 test, then crude odds ratios were calculated and reported for all studies in the set, when possible, to ensure comparability. Similarly, when studies within a metaanalysis set presented different types of ORs (dominant model vs additive model), ORs were recalculated to accommodate the measure that was common to all studies or calculable in a set. If a study reported only adjusted ORs without reporting crude numbers to recalculate ORs, adjusted ORs were used. The ORs for the metaanalysis were estimated using inverse variance weighted fixed effect models. All analyses were performed with STATA/SE 12.0 (StataCorp LP, College Station, TX).
Results
For the overarching topic of genetics and the genetic epidemiology of pelvic floor disorders, our literature search identified 423 nonduplicate articles, of which 125 met inclusion criteria based on the abstract, and 93 met inclusion after full text review. Of these articles, 21 pertained to the genetic epidemiology of POP ( Figure ). This included 1 GWAS, 2 linkage analyses, and 18 case-control candidate gene association studies involving 10 candidate genes (collagen type 1 alpha 1 ( COL1A1 ) (n = 5), collagen type 3 alpha 1 ( COL3A1 ) (n = 4), laminin gamma-1 ( LAMC1 ) (n = 3), matrix metalloproteinase 9 ( MMP9 ) (n = 3), matrix metalloproteinases 1 and 3 ( MMP1 and 3 ) (n = 2), lysyl oxidase-like 1 ( LOXL1 ) (n = 1), estrogen receptor alpha (ERα) (n = 1), estrogen receptor beta (ERβ) (n = 1), and progesterone receptor (PGR) (n = 1) ( Table 1 ). All studies were published in 2007 or later. Most had government and/or university grant funding, 71% (15/21).
| Author, year | Race and ethnicity (study country of origin) | Cases (with POP) | Controls (no POP) | Phenotype – Cases | Phenotype − Controls | Gene | SNPs evaluated a | Quality (Venice guidelines b ) |
|---|---|---|---|---|---|---|---|---|
| GWAS | ||||||||
| Allen-Brady et al | Cases: white (USA) Controls: white | 115 | Illumina iControlDB 2,976 | Treated for POP with a family history of prolapse or other pelvic floor disorders | Pelvic floor information not known. Excluded duplicate and closely related samples. | rs1455311, 4q21.21; rs1036819, 8q24.22; rs430794, 9q22.2; rs8027714, 15q11.2; rs1810636, 20p13; rs2236479, 21q22.3 | CBC (CCC for SNPs rs430794, rs1810636 due to lack of replication) | |
| Linkage analysis | ||||||||
| Allen-Brady et al | European descent (USA) | 70 cases (Familial study: 32 families) | N/A | Treated for moderate-severe POP, usually POP-Q stage III-IV (41/66) | N/A | LOD score 3.41; Chr9: 80.35Mb-88.81Mb with HLOD ≥1.86 | N/A | |
| Nikolova et al, | NR (USA) | 6 | Prolapse evaluated by POP-Q | LAMC1 sequence variant 1q31 | rs10911193 | N/A | ||
| Case-control, candidate gene analysis | ||||||||
| Chen et al, | White and African American; results reported by race (USA) | 165 (102 White 63 AA) | 246 (163 White 83 AA) | POP-Q stage III, IV | POP-Q stage 0-1; Cases and controls were matched on age, race, menopausal status, smoking history, BMI and parity. | LAMC1 | rs10911193; rs20563 | CCC |
| Chen et al | NR (Taiwan) | 69 | 141 | POP-Q stage II-IV | POP-Q stage 0-I | ERβ | rs2987983 (−13950 T/C) Promoter; rs1271572 (−12214 G/T) Promoter; rs9444599 (−1213 T/C) Promoter; rs1256049 (25652 A/G) Exon 6; rs1255998 (110943 G/C) 3′-UTR | CCB |
| Chen et al | NR (Taiwan) | 88 | 153 | POP-Q stage II-IV | POP-Q stage 0-I | ERα | rs17847075 (exon 1 C/T); rs2207647 (exon 1 G/A); rs2234693 (intron 1 T/C); rs3798577 (exon 8 C/T); rs2228480 (exon 8 G/A) | CCB |
| Chen et al | NR (Taiwan) | 84 | 147 | POP-Q stage II-IV | POP-Q stage 0-I | COL3A1 | rs1800255 (Exon 30 G>A); rs1801184 (exon 32 T>C); | CCB |
| Chen et al | NR (Taiwan) | 87 | 150 | POP-Q stage II-IV | POP-Q stage 0-I | PGR | rs500760 (exon 8 A/G); rs484389 (3′ untranslated region C/T) | CCB |
| Chen et al | NR (Taiwan) | 92 | 152 | POP-Q stage II-IV | POP-Q stage 0-I | MMP-9 | rs3918242; rs17576; rs2250889 | CCB |
| Cho et al | Korean (Korea) | 15 | 15 | POP-Q stage III-IV (women undergoing hysterectomy) | POP-Q stage 0 (hysterectomy for uterine myoma) | COLIA1 Sp1 binding site | No polymorphism seen at Sp-1 binding site in COLIA1 (all G/G, cases and controls) | CCC |
| Feiner et al | White or Ashkenazi-Jewish (Israel) | 36 | 36 | POP-Q stage III-IV POP | POP-Q stage 0-I | COLIA1 Sp1 binding site | SP-1 binding site (no rs#) | CCC |
| Ferrari et al | NR (Italy) | 137 | 96 | POP-Q stage II-IV | POP-Q stage 0-I | COL1A1, MMP1,3,9 | SP1 site of COL1A1 point mutation (G-T) in 1st intron; neg 1562 /T of MMP9 ; neg 1171 5A/6A of MMP3 ; neg 1607 1G/2G of MMP1 | CCB |
| Ferrell et al | African American and White (USA) | 137 | 141 | POP-Q stage II-IV | POP-Q stage 0-I, matched to cases on age, race, menopausal status, smoking history, BMI and parity | LOXL1 | No rs # (labeled -659 in promoter) | CCC |
| Jeon et al | Korean (South Korea) | 36 | 36 | POP-Q stage II-IV, Postmenopausal and parous | POP-Q stage 0-I, no stress urinary incontinence, postmenopausal and parous | COL3A1 | 5′-AAGTATACAAATTTCTAGATTG-3′ (forward)/5′-ATAAATGATCAGAAGGAAATCA-3′ (reverse) | CCC |
| Kluivers et al | European, Dutch (Netherlands) | 202 | 102 | POP present (not defined) | Vaginally parous, descent <1 cm above hymenal remnants, no prior POP surgery | COL3A1 | rs1800255 | CCC |
| Martinset al | White and Non white Brazilians (Brazil) | 107 | 209 | POP-Q stage III-IV, postmenopausal, no HRT | POP-Q stage 0-I, no documented vaginal surgery or stress incontinence, postmenopausal,no HRT | COL3A1 | No rs# (Labeled exon 31 G Allele) | CCC |
| Rodrigues et al | White and Nonwhite (Brazil) | 107 | 209 | POP-Q stage III-IV | POP-Q stage 0-I | COL1A1 Sp1 binding site | COL1A1 Sp-1 binding site polymorphism (no rs#) | CCC |
| Skorupski et al | NR (Poland) | 37 | 40 | POP-Q stage III-IV | POP-Q stage 0-I | COL1A1 | position 1240 in 1st intron; G -> T substitution; transcription factor Sp1 binding site of COL1A1 | CCB |
| Skorupski et al | NR (Poland) | 133 | 132 | POP-Q “grade” II-IV, undergoing surgery | POP-Q “grade” 0-I, dysfunctional uterine bleeding or undergoing TAH/SCH | MMP1,3 | MMP1 polymorphism (position −1607/−1608); MMP3 polymorphism (position −1612/−1617) | CCB |
| Wu et al | White, nonHispanic (USA) | 239 | 197 | POP-Q stage III-IV, not pregnant; no age cutoff but preferentially recruited younger women | POP-Q stage 0-I, no history of POP surgery, preferentially recruited older women | LAMC1 | rs10911193; rs1413390; rs20558; rs20563; rs10911206; rs2296291; rs12041030; rs12739316; rs3768617; rs2483675; rs10911211; rs41475048; rs1058177; rs12073936 | CCA |
| Wu et al | White, nonHispanic (USA) | 239 | 197 | POP-Q stage III-IV, no age cutoff but preferentially recruited younger women | POP-Q stage 0-I, no history of POP surgery, preferentially recruited older women | MMP9 | rs3918253; rs3918256; rs3918278; rs17576; rs2274755; rs17577; rs2236416; rs3787268 | CCA |
a Only SNPs with significant findings were reported for GWAS and linkage analyses. All evaluated SNPs are reported for candidate gene studies
b Venice guidelines grade (1) the amount of evidence, (2) whether replication was performed and (3) protection from bias.
The prolapse phenotype was most commonly defined by POP-Q stages II-IV, although some studies were more stringent, and 2 studies did not define the prolapse phenotype. All of the case-control studies defined the control as POP-Q stage 0 or 0-I. Many excluded women with connective tissue diseases. The GWAS and both linkage analyses were performed in families with a high rate of POP. All other studies were population based. Studies looked at Asian (33.3%, 7/21), European (23.8%, 5/21), and US white populations (23.8%, 5/21) ; 2 studies included subanalyses of African Americans ( ) and Brazilian white and nonwhite (9.5%, 2/21) populations ( Table 1 ). When reported, the mean or median age of prolapse cases ranged from 48 to 66 years and mean or median age of controls ranged from 49 to 69 years. Age was similar between cases and controls for 9 studies, had a discrepancy of ≥5 years for 6 studies, a discrepancy of ≥10 years for 3 studies and markedly disparate proportions of younger and older women in 2 studies (study did not report mean or median ages). Two studies preferentially recruited controls from an older population ; all other studies with an age discrepancy had controls that were younger than the prolapse cases.
Sample sizes were small across all of the studies (Venice category C, Table 1 ). The GWAS included 115 cases of POP and 2976 white controls from Illumina 550K (Illumina, Inc., San Diego, CA). The linkage analysis showing a predisposition for pelvic floor disorders on chromosome 9q21 involved 70 affected women from 32 families and mostly evaluated sister pairs. The linkage analysis showing an association with LAMC1 involved genotyping of 9 individuals from 1 family, 6 of whom had prolapse. Prolapse cases spanned 3 generations. Of the 18 candidate gene studies, sample sizes ranged from 15 to 239, with 9 studies having fewer than 100 cases. Control populations for these studies ranged from 15 to 246, with 5 studies having fewer than 100 controls. Only the GWAS had replication of its findings (Venice category B, Table 1 ). Methodology was strong for 2 papers and moderate for 8 studies (Venice categories A and B, Table 1 ).
A metaanalysis was performed on all SNPs evaluated by 2 or more studies and included type III collagen (a key component of connective tissue), matrix metalloproteinases (enzymes which degrade extracellular matrix proteins and likely play a role in tissue remodeling), and laminins (a component of the basement membrane involved in the structural scaffolding of tissue). By convention, each SNP is reported by its gene and a reference SNP identification number (rs#). If an identifier has not yet been assigned, then the location of the allele is listed. The reference and effect alleles have different nucleotide sequences which can be specifically stated (adenine [A], cytosine [C], guanine [G], thymine [T]).
Two studies evaluated each of the following genetic variants: COL3A1 rs1800255, MMP9 rs17576, MMP1 position −1607/8, and LAMC1 rs20563. Three studies assessed LAMC1 rs10911193. The metaanalysis suggests that COL3A1 rs1800255 genotype AA is associated with POP (OR, 4.79; 95% confidence interval [CI], 1.91–11.98; P = .001) compared with the reference genotype GG in populations of Asian and Dutch women. There was little evidence of heterogeneity for rs1800255 ( P value for heterogeneity = .94; proportion of variance because heterogeneity, I 2 = 0.00%). There was insufficient evidence to determine whether the following SNPs were associated with POP: MMP1 −1607/−1608 2G/2G (OR, 1.41; 95% CI, 0.85–2.33; P = .18) and MMP9 rs17576 GG/AG (OR, 0.87; 95% CI, 0.57–1.31; P = .50), LAMC1 rs10911193 TT/TG (OR, 1.16; 95% CI, 0.80–1.68; P = .43) and LAMC1 rs20563 of AA/AG (OR, 1.22; 95% CI, 0.88–1.68; P = .43) ( Table 2 ).
| Study | Effect | Ref | OR | 95% CI | Weight | P value a | Het P value b | I 2 c |
|---|---|---|---|---|---|---|---|---|
| COL3A1 (rs1800255) | ||||||||
| Chen | AG | GG | 0.74 | (0.41–1.26) | 40.38% | |||
| Kluivers | AG | GG | 0.96 | (0.59–1.58) | 59.62% | |||
| Metaanalysis | AG | GG | 0.87 | (0.59–1.27) | 100% | .46 | .52 | 0.00% |
| Chen | AA | GG | 4.59 | (1.17–18.05) | 44.98% | |||
| Kluivers | AA | GG | 4.95 | (1.44–17.06) | 55.02% | |||
| Metaanalysis | AA | GG | 4.79 | (1.91–11.98) | 100% | .001 | .94 | 0.00% |
| MMP1 (1607/1608) | ||||||||
| Ferrari | 1G/2G | 1G/1G | 2.24 | (1.16–4.30) | 42.05% | |||
| Skorupski | 1G/2G | 1G/1G | 0.96 | (0.55–1.67) | 57.95% | |||
| Metaanalysis | 1G/2G | 1G/1G | 1.39 | (0.90–2.09) | 100% | .15 | .05 | 73.40% |
| Ferrari | 2G/2G | 1G/1G | 2.81 | (1.25–6.33) | 44.98% | |||
| Skorupski | 2G/2G | 1G/1G | 0.93 | (0.49–1.75) | 55.02% | |||
| Metaanalysis | 2G/2G | 1G/1G | 1.41 | (0.86–2.33) | 100% | .18 | .04 | 77.50% |
| MMP9 (rs17576) d | ||||||||
| Chen et al | GG/AG | AA | 5.67 | (1.28–25.12) | 7.78% | |||
| Wu et al | GG/AG | AA | 0.74 | (0.48–1.14) | 92.22% | |||
| Metaanalysis | GG/AG | AA | 0.87 | (0.57–1.31) | 100% | .5 | .01 | 84.90% |
| LAMC1 (rs10911193) d | ||||||||
| Chen et al African Americans | TT/TG | GG | 1.83 | (0.59–5.65) | 10.85% | |||
| Chen et al Whites | TT/TG | GG | 0.88 | (0.48–1.62) | 37.68% | |||
| Wu et al | TT/TG | GG | 1.29 | (0.77–1.68) | 51.47% | |||
| Metaanalysis | TT/TG | GG | 1.16 | (0.80–1.68) | 100% | .43 | .46 | 0.00% |
| LAMC1 (rs20563) d | ||||||||
| Chen et al African Americans | AA/AG | GG | 1.43 | (0.56–3.65) | 11.78% | |||
| Chen et al Whites | AA/AG | GG | 0.8 | (0.45–1.46) | 29.19% | |||
| Wu et al | AA/AG | GG | 1.44 | (0.95–2.19) | 59.03% | |||
| Metaanalysis | AA/AG | GG | 1.22 | (0.88–1.68) | 100% | .23 | .28 | 22.30% |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
