Varicose Veins as Model for Apportionment among Risk Factors for Compensation Purposes

The National Insurance Institute (NII) is responsible for evaluating all occupationally related injury in Israel. The NII receives approximately 90,000 work-related claims annually and must determine if an injury is indeed work related. Furthermore, if there are other causes, such as a pre-existing disease or non-work related factors, then the NII must apportion compensation among these causes. The process that is utilized by the NII to legally establish causation for occupationally related injury or disease is based on case law (the Sitruk Statute). This statute requires that in order to attribute causation, two criteria must be met. The first criteria is satisfied when the claimant is able to prove with greater than 50% certainty that the injury or disease can be a direct result of a particular work condition. In other words, it must be “more likely than not” that an exposure or work circumstance is the cause of the disease. There is no quantitative measure to address the court's requirement of 50% certainty; it is a qualitative assessment. If this can be demonstrated, then the second criteria will be satisfied if it can be proven that at least 20% of the injury or disease can be directly attributed to the exposure or circumstance (attributable risk). In the event that the “50/20” threshold is reached then the NII will compensate the claimant.

Historically, this process has been “informal” and expert driven. Recently, the NII has introduced a more formal process. The basis for this determination by the NII is an extensive literature review to determine what risk factors exist for the relevant disease and then apportioning among the factors present in each claim. The volume of claims and the short time allowed to process the claim precludes a formal meta-analysis for each disease. This paper will describe this process using varicose veins as an example.

Chronic venous disease (CVD) is a well-recognized phenomenon affecting the general population. In addition to its cosmetic aspect, CVD affects quality of life as well¹  causing pain and reducing work productivity, resulting in significant financial burden. CVD prevalence is highest in Western countries, consuming 2% of healthcare budgets² . Varicose Veins (VV) are one of the clinical presentations of CVD³  and range in severity from being cosmetic in nature to exhibiting significant symptoms. Their prevalence is high, with epidemiological studies describing ranges of 2%-56% in men and <1%-60% in women⁴ . A number of potential risk factors have been suggested, with prolonged standing among them⁵ . Aside from the medical significance, since many countries compensate their workers for occupationally related disease, the import of prolonged standing is of particular interest. There is no uniform definition of prolonged standing. Since it is highly unusual to stand motionless (static standing) for an entire work shift, it will be important to define what prolonged standing is.

VV can be defined as twisted, dilated, palpable, subcutaneous veins, generally greater than 3mm in diameter⁶ . The exact mechanism for the development of VV is not clearly defined⁷ ; however, the main factors in the etiology of varicose veins are venous dilation and valvular insufficiency⁸  while central venous obstruction may also contribute. In 1994, the American Venous Forum devised the CEAP (Clinical, Etiologic, Anatomic, and Pathophysiologic) classification in order to allow for a uniform description of CVD. CEAP classification runs on a scale of C0 (no signs of venous disease) to C6 (active venous ulcer), with VV being classified as C2⁹ .

Using PubMed, a literature search was conducted for risk factors associated with varicose veins (Fig. 1). The key words “varicose veins” were entered in association with “standing” and “risk factors”. The term “varicose veins” was also entered with “occupation”, “job”, and “work”, respectively. Additionally, “chronic venous disease” was entered along with “standing”. After an initial search, a secondary literature search was performed to include the following specific potential risk factors in the analysis: Age, Family History, Gender, Smoking, Pregnancy, Body Mass Index (BMI), Height, Deep Vein Thrombosis (DVT), Oral Contraceptive use, and Prolonged Standing. By conducting both searches, 9,036 articles were identified. Filtering for English language, human studies, full text articles, and publications dated between 1980 and 2019, yielded 2,356 studies. Articles that addressed epidemiology and risk factors were selected (rather than pathophysiology, genetics, diagnosis or treatment), resulting in 96 relevant studies. Articles on studies, which related the particular risk factor solely to disease progression, ulcers or clinical severity, were also excluded. After removing the excluded articles and duplicates, 30 relevant articles were identified. Utilizing the “Similar Articles” tool found in PubMed, in addition to bibliographical references, including those found in review articles, 32 additional studies were identified. Unless the study made specific reference to varicose veins, data from studies on chronic venous insufficiency (CVI) was excluded, as the term may refer to CEAP levels greater than C2. Therefore, an additional 12 articles were removed, as they did not specifically refer to varicose veins but rather to CVI, CVD or multiple CEAP levels.

A qualitative score from A to D (with A being the highest score) was assigned to each study by MW and MT, and, when necessary, consensus was reached. The following study characteristics were taken into account in the rating system for quality:

1. Study type, where longitudinal or prospective studies had greater weight than cross-sectional or retrospective studies.

2. Sample size of the outcome of interest.

3. Studies in which patients underwent physical examination and/or ultrasound were given greater weight than those by questionnaire only.

4. The use of CEAP classification was given greater weight.

5. Limitations and or bias were addressed in the study.

6. Studies with multivariate analysis results were given greater weight.

7. Studies that included quantification (e.g., hours spent standing, BMI, number of pregnancies) were given greater weight.

Impact Factors (IF) were compared with these qualitative rankings. IF's were obtained from SciJournal.org when available.

Data from articles that referred to surgery, CVI or used duplicate populations were not combined. Because cohort and case control studies were combined, sample sizes were taken to be the population diagnosed with venous disease in each respective study, rather than the entire study population, as the study population size of the cohort studies is typically many times larger than that of the case control studies.

An attempt was made to synthesize the available data through meta-analysis, however the paucity of data made this unfeasible. The single meta-analysis in this review dealt with pregnancy as a risk factor for VV¹⁰ . The OR from that study was not combined with other data, as data from the individual articles in that study were included in the overall analysis. Data was stratified by gender when available. Relative risks or odds ratios for male and female genders were typically not combined, however some studies produced pooled data only. “Dose response” was evaluated for age, pregnancies, BMI, and prolonged standing. When the number of pregnancies was reported as ≥1 they were considered as one pregnancy for calculating the weighted risk.

In the final analysis, 50 original research articles dealing with adult human subjects that evaluated which factors may contribute to the development of C2 varicose veins were identified. Of the 50 research articles, 32 were cross-sectional studies, 10 were longitudinal/prospective, 6 were case control, 1 was a retrospective cohort, and 1 study was a meta-analysis^*. A table of all the studies (Supplement) was constructed to include study type, population studied, risk factor correlation to varicose veins, RR/OR/HR for prolonged standing, study ranking, and journal IF (with year of publish and most recent). On average, IF's of studies which were considered to be of higher quality were greater than for those of lower quality (Table 1).

Twenty-four studies adequately fulfilled our criteria of higher quality (Table 1). Of the ten risk factors which were studied, eight showed a positive correlation with the presence of varicose veins, and were felt to fulfill the 50% criteria of being more likely than not. Thus, smoking and oral contraceptive (OC) use were excluded as risk factors, as many studies did not show a statistically significant correlation with developing varicose veins (Table 2). A number of studies did not contain relative risks or odds ratios. Some studies utilized duplicate populations, while others combined CVI and varicose vein data to produce a single relative risk. Therefore, eleven remaining studies contained appropriate data that could potentially be combined. Each article was grouped by its corresponding risk factor or factors. A summary table was then constructed, which outlined the weighted relative risk of each individual risk factor (Table 3).

For each risk factor, a summary RR was generated. The summary RR for prolonged standing was found to be 1.37 for females and 1.51 for males. The seven other factors revealed significantly elevated risks as well. For example, our calculated risk for family history was 2.03 in females and 2.07 for males. Female gender had an elevated risk of 2.14. A linear relationship to the presence of varicose veins was also found with respect to age and number of pregnancies.

where P_e represents the proportion of subjects being exposed in the population¹¹ .

Here, since 100% of the members of the population have all the particular risk factors, P is equal to 1, and mathematically falls out (as above).

Using the RR from the literature survey allows the calculation of attributable risk for all the risk factors present, under the assumption that the risks are additive. The mathematical formula for AR for a specific factor (AR₁) is shown above, where N equals the number of risk factors to which the individual has been exposed. It differs from population attributable risk, which attempts to estimate the contributions of particular risk factors in a population where the prevalence of risk factors varies. When calculating attributable risk for populations, RR-1 is multiplied by the prevalence of a particular risk factor in the population (P).

In order to utilize the above formula, one enters the RR for the particular factors present in a specific patient/claimant from the risk factor table, and using the spreadsheet (available from authors) the AR of each factor is calculated. In this way, the respective contribution of each risk factor is determined as a percentage of the total. A baseline risk in the population is taken into account as well, thus the combined apportioned risk factor percentages always equals 100%. The RR for female gender is not applied when the data is already stratified by gender but rather only in cases where pooled data is used. This apportionment allows us to determine if the second court criteria of having an attributable risk of greater than 20% is satisfied.

In the above example, the person had two pregnancies (RR=1.73) and prolonged standing at work (RR=1.37). Thus, the AR for pregnancy was 35%, the AR for prolonged standing was 17%, and the remaining baseline population risk was 48%. In this case, the 20% threshold with respect to prolonged standing was not attained.

In the example below, a 55-year-old male security guard developed varicose veins, which he attributed to prolonged standing at work. The medical history revealed an elevated BMI. RRs are assigned as weight 1.23, age 1.24, and prolonged standing 1.51. The calculated AR for standing is 26%, which surpasses the disability threshold.

Apportioning among risk factors is a critical need for occupationally related diseases. Using prolonged standing and varicose veins as an example, the above process describes a rapid approach for many diseases. Prolonged standing position at work is often cited in the medical and lay literature as one of many possible etiological factors in the development of varicose veins. The proposed mechanism described by the Canadian Centre for Occupational Health and Safety is that prolonged and frequent standing, without some relief by walking, causes blood to pool in the legs and feet¹² . On the other hand, a number of studies have found no significant association^13–15 . In this review, ten possible risk factors for varicose veins were identified, prolonged standing being just one of them. A recent systematic review similarly explored the association between these risk factors and the presence of varicose veins¹⁶ . Since the NII is charged with compensating work-related injuries based on apportionment, we incorporated a method to more accurately quantify the attributable risk for each factor for each claimant.

A number of limitations were encountered in summarizing the literature. Standing is rarely measured prospectively as most of the studies reviewed in the literature are cross-sectional or retrospective. As noted in previous studies, occupational exposures were primarily obtained through self-report and job titles¹⁷ . The studies which were reviewed relied heavily on physician documentation in the medical records such as an appropriate history and physical examination or ultrasound evidence to corroborate the diagnosis. As not all studies utilized the uniform CEAP classification, comparing and combining data was not always possible. Although the majority of studies that were reviewed do show an association of prolonged standing with the presence of VV, the findings are inconsistent. Similarly, there is disagreement among many studies regarding other risk factors as well. Furthermore, the unit of measure was frequently different. Some studies included data as “per decade” when evaluating age as a risk or “per pregnancy” when evaluating pregnancy as a risk factor. Therefore, combining results from such studies was not possible. Similarly, most studies lack “dose response” curves, making the assignment of causality more difficult. In addition, not all studies reviewed include odds ratios or relative risks. Finally, the results of these studies are to be taken with caution as they may be subject to recall and misclassification bias.

The heterogeneity between the studies limited how the data could be combined. Due to these limitations in combining the data, additional methodologies were considered as well. For example, unweighted relative risks for standing were combined which included all studies regardless of their qualitative score. The results were quite similar to the results presented in the current study. Limiting the risk table to combining data only from those studies that were statistically significant had little effect on the results.

The evaluation of “standing” is also difficult due to the imprecise definition and quantification of prolonged standing, both in terms of hours standing per day of work and duration of exposure over time (in years). Studies that mention years spent in occupations that require a standing position generally indicate the average number of years in employment without directly relating it to disease development. While some studies assessed 4-6 hours of standing per workday, they rarely assessed time spent in static standing only. The literature review also found studies which combined standing with sitting or walking. The studies which were reviewed did not all use the same cut point for the number of hours to be considered as prolonged standing. Our study considered the term “prolonged standing” as 75%-80% of the workday. Based on the data that was available, the determination of risk regarding 4-6 hours of standing could not be estimated, either because the data was not present in the studies that were used in our final analysis or because statistical power was lacking. Our comprehensive risk table therefore reflects risk for standing at least six hours per workday. In our review of the medical literature, research was not found to address the issue of latency of exposure with respect to prolonged standing, i.e. time elapsed from exposure to development of symptomatology.

A strong body of evidence indicates that heredity plays an important role in the etiology of VV¹⁸ . The present review and analysis of the data is in agreement with those beliefs as the literature suggests a positive correlation with family history of VV. The role of heredity has been demonstrated in family and adoptee studies as well^19–20 . Twin studies indicate that venous function is influenced by genetic factors. In one study, venous compliance and capacity were correlated in monozygotic twins but only weakly in dizygotic twins²¹ . Another study demonstrated a 90% risk for VV if both parents were VV positive²² . It is well recognized that CVD runs in families, nevertheless, the genetic basis for this connection is yet to be elucidated²³ . There appears to be a very significant genetic contribution to the development of VV, possibly stemming from mutations in the FOXC2 gene²⁴ . In one study, pathological reflux was found on duplex ultrasound in all 18 participants with a FOXC2 mutation²⁵ . A large genome wide association study identified 30 genetic variants associated with varicose vein disease²⁶ . Interestingly, studies also show increased expression of estrogen receptors in veins of patients with CVD^27–29 . Our review likewise found increased risk for female gender and for parity.

Varicose veins are a significant healthcare issue in the general and working population. There appears to be a very strong genetic component, with other risk factors contributing as well. Numerous risk factors have been investigated. This paper describes how the NII assesses causality by attempting to quantify the attributable risk for each risk factor that is present in an individual claim. In the past, for the most part, an “experience based” approach was used to assign apportionment. By constructing a weighted comprehensive risk table, our results allowed us to more accurately apportion among the various risk factors present in individual claims and compensate them accordingly. Of course, there is room for applying individual judgement to each case. This approach is also being applied to other diseases and their risk factors.

Most of the papers in this field consisted of retrospective, cross-sectional studies, which resulted in significant limitations. Duration and degree of exposure was for the most part self-reported by patients in these studies. Nevertheless, our results suggest that prolonged standing at work may play a role in the development of varicose veins. Our findings suggest that for most people with VV, the attributable risk for standing will be relatively low, as other factors typically contribute to the overall risk. Our methodology allowed us to reach a claim-specific determination of causality and apportionment. At the NII, each case is evaluated individually, based on their respective risk factors, and our factor-specific estimate of risk. Thus, apportionment can be used for any occupationally related disease.

The authors acknowledge the difficulty in combining data from multiple studies due to the likely presence of heterogeneity. In addition, the disparate methodologies of the various studies compounded that difficulty. Nevertheless, because of the many claims, which need to be resolved by the NII in a timely fashion, an evidence based approach, relying on the accepted medical literature was sought. While the proposed method has its limitations, it may serve as a practical tool for the resolution of these claims.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

None