Thromboembolic Events Are Increased After Splenectomy in Postmenopausal Women

Abstract

In the United States, 22 000 splenectomies are performed annually [1] for various conditions. Splenectomy has been linked to overwhelming sepsis and more recently to thromboembolic events (TEE) [2-5]. Danish national registry studies documented the increased occurrence of TEE postsplenectomy; however, earlier studies were restricted to males and later ones, including two Danish studies, did not discriminate TEE in women from those in men [2-4]. TEE risk differs between men and women in several ways [5], so sex-specific assessments are critical. Premenopausal women have a predisposition to TEE for several reasons including hormonal and autoimmune ones; however, TEE risk remains unclear in postmenopausal women. We therefore conducted a prospective cohort study within the Nurses' Health Study (NHS) using biennial questionnaires with medical record confirmation of thromboses to robustly assess association between splenectomy and TEE in postmenopausal women.

The NHS was established in 1976 and enrolled 121 700 female registered nurses between 30 and 55 years old. Splenectomy history was included on the 2004 questionnaire and 90 853 postmenopausal women (at that time between ages of 58 and 83 years) were included in the current analysis. The biennial questionnaires asked participants whether they had stroke, myocardial infarction (MI), deep vein thrombosis (DVT), or pulmonary emboli (PE) since cohort inception. Participants self-reporting TEE received letters requesting confirmatory medical records. Through 2012, questions about DVT and PE were distinct; subsequently, 2012–2016, participants were asked about both as a single outcome.

The primary outcomes included DVT, PE, stroke and MI, self-reported on biennial questionnaires. Follow-up for this nested cohort study started in 2004 after all splenectomies. End of follow-up was through the 2016 questionnaire, date of death, or diagnosis of outcome, whichever came first. Loss to follow-up in NHS is < 5% every 2-year cycle. Date and cause of death were obtained through linkage with the National Death Index. We conducted a prospective analyses (start from 2004) with Cox Proportional Hazard models to generate hazard ratios and 95% confidence intervals (CIs). We included the following covariates as potential confounders in multivariate models: (a) continuous variables: age, BMI, hypertension, and cholesterol level; and (b) yes/no variables: smoking status, physical activity, diabetes, and menopausal hormone therapy.

Biennial questionnaires did not include a specific question about reason for splenectomy. Participants had the option to report these diagnoses under “other illnesses.” They were then mapped to ICD-9 and ICD-10 disease codes including categorizing these conditions into definite ITP (codes = 287.31), and possible ITP (codes = 287.30). Thirty-four participants were considered “ITP,” too few to perform a sub-analysis. The reason for most splenectomies was unknown.

In 2004, 323 participants in NHS reported having had a splenectomy, among the 90 853 participants completing that year's questionnaire. The mean age of study participants was similar for those reporting splenectomy (64.5 years) and those without splenectomy (63.6 years) (Table 1A). Participants reporting a splenectomy were similar to women who never underwent splenectomy on height, BMI, race, comorbidities (e.g., diabetes, high blood pressure, elevated cholesterol) and covariates (e.g., aspirin, smoking status, past physical activity) (Table 1A). Participants were all female and overwhelmingly white consistent with demographics of those entering nursing in 1976. Seventy-five percent of participants who underwent splenectomy did so in 1996 or before, 14% in 1997–2001, 2% in 2002, 3% in 2003, and 2% in 2004.

In age-adjusted models (Table 1B), women who had undergone splenectomy had a greater than threefold risk of DVT compared to women who had not (hazard ratio 3.16 [95% CI: 1.50–6.66]). In the fully multivariate model, the HR was 3.19 (95% CI: 1.51, 6.71) for TEE in splenectomized women. The impact of different variables (confounders and comorbidities) was negligible. Similarly, women with splenectomy had 3.7-fold risk of PE compared to women without splenectomy (HR multivariate = 3.70 [95% CI 2.04, 6.71]). When considered together as a composite outcome, splenectomized patients had a 2.82-fold higher risk (95% CI: 1.81–4.38) of VTE (DVT and/or PE) compared with those not having undergone splenectomy (Table 1B).

The multivariate-adjusted HR of ATE for splenectomized women was 1.77 (95% CI: 0.79–3.95). Both stroke 1.77 (95% CI: 0.79–3.95) and MI (HR 1.72; 95% CI: 0.71–4.14) tended to have increased incidences after splenectomy but CIs crossed unity (Table 1B). Event numbers were small.

Among splenectomized women, 34 (10.53%) self-identified as having ITP. In the non-splenectomized cohort, 109 (0.12%) had ITP, too few to do a sub-analysis. Self-reported hemolytic anemia was substantially less frequent in both groups. The reason to perform splenectomy was unknown in most cases.

In this NHS analysis, splenectomy was associated with a substantially increased incidence of VTE in postmenopausal women (no men participated in NHS). After adjusting for multiple covariates and potential confounders, splenectomized women had a 2.7–3.7-fold higher risk of DVT, PE and the composite DVT/PE (VTE) endpoint than controls. This is the first large-scale, longitudinal, prospective analysis of TEE specifically of postmenopausal women after splenectomy and it demonstrated a strong association between splenectomy and VTE in these women. Given that women in NHS were postmenopausal, they did not have prothrombotic influences seen in younger women: neither endogenous endocrinologic effects nor higher prevalence of autoimmune disease.

In this study, we identified an approximately threefold increased risk of VTE (DVT, PE, and DVT/PE) after splenectomy in our unique study population of predominantly white, female nurses with an average age in 2012 over 70 years. The incidence of VTE has been reported to increase in women after the age of 70 years [6]; however, in this NHS study, none of many potential confounders, including age and being on hormonal therapy. were identified as significantly affecting the findings in Table 1B. As with other population-based studies, while this association does not prove causality, it suggests that splenectomy increases the risk of VTE in postmenopausal women.

For ATE, the risk of stroke and MI were only borderline increased in the splenectomized women. The findings were not significant largely because the overall incidence of stroke and MI were low. In population-based studies, not specific to splenectomized patients, cardiovascular disorders have later onset in women [5]. The risk of cardiovascular diseases is related to hormonal levels: women were protected from cardiovascular diseases before menopause, but their risk of such events increased after menopause. This could be explained by increased platelet activation in postmenopausal women [7, 8].

Previous studies describing splenectomy-associated TEE risks often focused entirely on males, or did not report gender-specific data. A study of hereditary spherocytosis (HS) patients showed a 5.6-fold higher arteriosclerotic events incidence postsplenectomy. HS females undergoing splenectomy before age 18 had a 22% cumulative ATE rate by age 70, lower than that of males, 32%.

They also had later first-event onset (58–83 years) than did males (46–80 years) [9]. The risk of VTE did not differ by sex [10]. These small cohorts in a prothrombotic disease remain the only previously published study providing specific information on postsplenectomy TEE and VTE specifically in women.

Explanation of the etiology and degree of VTE post splenectomy in postmenopausal women still requires explanation. Our study makes clear that the increased risk of DVT/PE in postmenopausal women postsplenectomy is not explained by age, hormonal therapy, or other factors. The increased risk of VTE after splenectomy may influence clinicians to implement thromboprophylaxis, withhold prothrombotic medications (e.g., postmenopausal hormones), or perform diagnostic testing for risk of VTE at a lower threshold in postmenopausal women after splenectomy.

Particular strengths of our study are the high quality of self-reported health data by almost 100 000 health care professionals (nurses) and very high follow-up rates in the NHS. TEEs were validated by medical record review and have been shown to be accurately self-reported by patients in studies, including in this population. Uniform gender and age of the participants provides clearer findings with less intrinsic variability.

Limitations include lack of family history; whether there was thromboprophylaxis; not distinguishing immediate and delayed postsplenectomy TEEs; the small event numbers with restricted observation period and age range; and absence of laboratory investigations and splenectomy indications or underlying diseases. Additional follow-up could not be obtained because of advanced participant age.

This study showed for the first time in postmenopausal women that splenectomy clearly increased the risk of VTE (DVT, PE) 2.7–3.7 fold and, less so, ATE (MI and stroke) 1.7–1.8 fold. In the decision whether to proceed with splenectomy in a woman, the long-term increased risks of VTE and possibly also ATE, as well as of sepsis, must be carefully balanced against the efficacy of splenectomy for disease control in addition to the availability, efficacy, toxicity, and cost of alternative treatments; especially since splenectomy may likely be the less expensive option. Finally, as splenectomy appears to increase the risk of TEE in postmenopausal women, we suspect that the TEE risk postsplenectomy may be even further increased in premenopausal women due to high hormone levels and autoimmune contribution to VTE risk.

J.B.B. and R.M.T. designed the study and performed research. C.K. Y.M., Y.C., and B.R. analyzed data. Q.F. and J.B.B. wrote the paper. All authors reviewed the data, read and edited the manuscript.

The authors have nothing to report.

This study was conducted in accordance with the ethical principles of the Declaration of Helsinki.

Written informed consent was obtained from all participants prior to their inclusion in the study.

The authors declare no conflicts of interest.