Introduction

Low-molecular-weight heparins (LMWHs) are anticoagulants made by depolymerization of unfractioned heparin (UFH) [1, 2]. In the last decades, LMWHs have largely replaced UFH as anticoagulants, because they are at least equally effective in prevention and treatment of venous thromboembolisms (VTEs) and have many practical advantages, most importantly the possibility of subcutaneous administration without the need of routine laboratory monitoring of the anticoagulant response [3–8]. Moreover, LMWHs have a more predictable anticoagulant response, longer half-life (allowing once or twice daily administration), and a dose-independent elimination and they cause less heparin-induced thrombocytopenia [1, 2, 9, 10]. There is conflicting evidence whether bleeding complications differ between LMWHs and UFH, although a Cochrane analysis showed that LMWHs reduce the occurrence of major bleedings [8, 11–14].

The molecular weights of the various LMWHs differ. Tinzaparin has the highest average molecular weight (6500 Da), while certoparin has the lowest average molecular weight (3800 Da) [13]. Pharmacokinetics of different LMWHs vary: they differ in elimination half-life, clearance, and bioavailability [15, 16]. This might be a consequence of the differences in molecular weight. Also with regard to pharmacodynamics, LMWHs vary: anti-Xa/anti-IIa activity ratios range between 1.5 and 2.5 (tinzaparin and certoparin) to 3.6–6.5 (reviparin) [13].

Since LMWHs are mainly excreted by the kidney, they may accumulate in patients with renal insufficiency increasing the risk of bleeding [17–19]. Due to above-described differences in pharmacokinetics, data on accumulation in renal insufficiency cannot be easily converted from one LMWH to another [2]. For instance, therapeutic dosages of different LMWHs do not all accumulate in patients with renal insufficiency: nadroparin and enoxaparin were found to accumulate, while tinzaparin did not [20–26]. For LMHWs that accumulate in a therapeutic dosage, dose reduction is often recommended. Also for the lower prophylactic dosages, such dose reductions have been suggested, especially for high prophylactic dosages such as those used in cancer patients or in high-risk surgery [24, 27–29].

Even in this era of novel oral anticoagulants, LMWHs will stay important for the initial treatment of VTE during the first 5–7 days in patients treated with dabigatran and edoxaban, as well as for prophylaxis of VTE. Especially for patients with severe renal insufficiency in which case the novel oral anticoagulants are contraindicated, it remains important to know whether LMWHs can be prescribed safely.

Therefore, we conducted a systematic review to investigate whether prophylactic dosages of various LMWHs lead to accumulation defined as an increase in anti-Xa activity and whether accumulation depends on molecular weight of the LMWH.

Materials and methods

The meta-analysis was prepared in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement [30]. No prespecified formal protocol was registered.

Article search

A comprehensive, systematic literature search has been conducted on 17 February 2015 using Embase, Medline, Web of Science, Scopus, Cochrane, PubMed publisher, and Google scholar. The syntax emphasized for LMWHs, impaired renal function, and pharmacokinetics using synonyms and relevant terms. The search terms as used in Embase, as an example, are shown in a supplemental file. In addition to the search results, we manually searched reference lists of all relevant articles for additional studies.

Inclusion criteria and exclusion criteria (eligibility criteria)

We only included articles in English on prophylactic LMWH treatment, studying at least ten patients (non-pregnant and with no children) with renal impairment. We defined accumulation as an increase in anti-Xa activity after consecutive administration for several days. Therefore, we excluded studies which did not administer LMWH on consecutive days (except for studies in patients receiving renal replacement therapy) and did not measure anti-Xa activity on multiple days while giving LMWH. We excluded case reports, overviews, expert opinions, recommendations, reviews, and replies on articles. Abstracts of unpublished data were not excluded; authors were approached for additional information.

Study selection and data collection

A single reviewer excluded articles that did not meet the eligibility criteria by using the title and abstract. If necessary, the full text was read. If the reviewer could not decide whether to exclude the article, a second reviewer was asked for advice and both met to reach a consensus. Furthermore, a single reviewer collected relevant data from the included articles for the review.

Risk of bias

Risk of bias was evaluated at study level using the Cochrane Risk of Bias assessment tool. The criteria on random sequence generation, allocation, and blinding were disregarded.

Results

Study selection

The search yielded 1387 articles from which 713 were duplicates. Figure 1 shows a flow diagram illustrating literature evaluation.

Fig. 1
figure 1

Flow diagram illustrating literature evaluation

Full size image

Characteristics

Table 1 shows the study and patient characteristics. We included 11 articles of which one was on LMWH in continuous venovenous hemofiltration (CVVH) and one on hemodialysis patients. Five studies examined dalteparin accumulation, and five articles examined enoxaparin [31–37, 40, 41]. For tinzaparin, bemiparin, and certoparin, there was one study for each LMWH [35, 38, 39].

Table 1 Study characteristics of included studies including patient characteristics and outcome measurements
Full size table

All studies were conducted prospectively. Two were randomized trials, and eight were cohort studies. Only one study did report clear prespecified primary outcomes. The number of patients ranged between 12 and 138. Three studies only included patients on an intensive care unit (ICU), and three studies only included elderly patients (Table 1). The follow-up length per patient was 4 days to 3 weeks. Accumulation was mainly assessed by measuring peak anti-Xa activity or trough anti-Xa activity (Table 2). Six studies reported clinical outcomes like hemorrhagic events, thrombosis, and mortality rates.

Table 2 Outcomes of included studies
Full size table

Dalteparin

Dalteparin accumulation was studied in five articles. In patients on a general medical or surgical ward, no accumulation was found in a total of 157 patients by measuring peak anti-Xa activity on day 6 or day 10 [32, 34].

At the ICU, trough and peak anti-Xa activity did not show accumulation on day 9 or day 12 in a total of 157 patients not undergoing CVVH [31, 33].

Also in hemodialysis patients (n = 7) prescribed 2500 IU dalteparin during hemodialysis sessions for 4 weeks, anti-Xa activity at different time points (1, 2, 3, 4, 24, and 28 h postinjection) were not significantly different between week 1 and week 4 [40]. The patients underwent hemodialysis sessions three times a week for 4 h.

Enoxaparin

Three studies of which two in patients older than 75 years examined enoxaparin accumulation in, respectively, 125, 28, and 48 non-dialysis patients [35–37]. The studies in the elderly with CrCl 20–50 ml/min found a significantly higher maximum concentration (C max), area under the curve (AUC), and trough anti-Xa activity (24 h postinjection) after 8 days administration and a higher peak anti-Xa activity after 10 days, compared to patients with better renal function [35, 36]. The other study in patients with different severity of renal insufficiency found accumulation (a significantly higher C max, longer half-life, higher AUC) already on day 4 in patients with CrCl ≤ 30 ml/min [37].

Hemodialysis patients (n = 7) prescribed 40 mg enoxaparin at hemodialysis sessions for 4 weeks, anti-Xa activity at different time points postinjection were not significantly different between week 1 and week 4 [40]. Also in ICU patients (n = 12) undergoing CVVH with a flow rate of 30 ml/kg/h, no accumulation of enoxaparin (30 mg daily) was found [41].

Tinzaparin, bemiparin, and certoparin

Administration of 4500 IU tinzaparin for 8 days in 27 patients older than 75 years with CrCl 20–50 ml/min and body weight below 65 kg did not lead to significant changes in C max, AUC, and trough anti-Xa activity (24 h postinjection) [35].

For 3500 IU bemiparin, C max was higher and mean half-life was 2–4 h prolonged in patients with severe renal insufficiency after 4 days administration in 48 patients [38]. AUC significantly increased with the degree of renal insufficiency.

For 3000 IU certoparin, C max and AUC were significantly higher in patients with renal insufficiency compared to healthy controls after 5 days administration in 24 patients [39].

Adverse events

Six studies reported VTEs and bleeding events, although all were underpowered to find significant correlations [31–36]. In three of these studies, anti-Xa activity was undetectable during bleeding [31, 33, 34]. Five serious bleeding complications were reported when using enoxaparin, but anti-Xa activity in these patients was the same as in those without bleeding (p = 0.77) [36].

Discussion

In this systematic review, we show that prophylactic dosages of dalteparin and tinzaparin did not accumulate in patients with renal insufficiency, while prophylactic dosages of enoxaparin, bemiparin, and certoparin did accumulate. Dalteparin also showed no accumulation in hemodialysis patients. Surprisingly, enoxaparin did not show accumulation in hemodialysis and CVVH patients, which might be due to removal by renal replacement therapy [42]. No data are available for nadroparin in patients with renal insufficiency.

These results are in accordance with studies on therapeutic dosages, except for dalteparin. Dalteparin is the only LMWH that appears to accumulate when used in a therapeutic dosage, while in studies using prophylactic dosages no accumulation was detected [31–34, 43]. A single-dose study showed reduced elimination of prophylactic dalteparin in patients with renal insufficiency, but apparently clearance is sufficient to prevent accumulation [31–34, 44].

Data on assessment of accumulation on prophylactic nadroparin in renal insufficiency are lacking. Only one multiple-dose study in six patients with CrCl above 30 ml/min and one single-dose study were conducted [45, 46].

Our findings confirm the theory that accumulation seems to depend on the mean molecular weight of LMWHs as shown in Table 3. LMWHs with the lowest molecular weight (enoxaparin, bemiparin, certoparin, and nadroparin) all showed accumulation in a therapeutic dosage and a prophylactic dosage. Tinzaparin (the LMWH with the highest mean molecular weight) has shown not to accumulate in neither therapeutic nor prophylactic dosage. The most likely explanation for this counterintuitive relationship between size and accumulation is that the larger molecules are less dependent on renal clearance [4, 28, 49, 50].

Table 3 Accumulation dependency on molecular weight
Full size table

In patients undergoing renal replacement therapy, it has been shown that many LMWHs are safe to use even in therapeutic dosages [28, 51]. For assessment of accumulation, only a few studies have been conducted. Most of these studies were excluded from this review as they included a single bolus administration, a therapeutic dosage, or no anti-Xa activity measuring at multiple days. Dalteparin showed accumulation in hemodialysis patients if prescribed in a therapeutic dosage, but not in a prophylactic dosage, whereas accumulation was found for prophylactic dosage in peritoneal dialysis patients [19, 40, 52]. Prophylactic dosages of enoxaparin and prophylactic and therapeutic dosages of nadroparin showed no accumulation in hemodialysis and CVVH patients [40, 41, 53, 54]. Tinzaparin accumulation has been found in a therapeutic dose in hemodialysis patients, but in a prophylactic dose, anti-Xa activity returned to baseline in 24 h [52, 55]. In conclusion, for dalteparin accumulation in renal replacement therapy is comparable to accumulation in patients without renal replacement therapy, but in other LMWHs accumulation in renal replacement therapy seems not to depend on mean molecular weight of LMWHs. The mechanism for accumulation in renal replacement therapy is unknown, but the highly negative charge of LMWHs might play a role [52]. More studies are needed to assess LMWH accumulation in patients on renal replacement therapy.

The strength of this review is first that we included only articles that objectively observed accumulation based on anti-Xa activity rather than an accumulation prediction based on the half-life or the time it takes for anti-Xa activity to return to baseline. Second, by including the recent studies on bemiparin and certoparin, this paper confirms the earlier stated theory that accumulation depends on the mean molecular weight of LMWHs.

A limitation might be that we did not include single-dose studies. However, a single-dose study is not suitable for objectively detecting accumulation [28, 32]. A significantly lower clearance or prolonged half-life for a LMWH in a single-dose study does not necessarily indicate that the LMWH accumulates, as is clear from the findings on dalteparin. Furthermore, follow-up of patients in some studies was relatively short (in some studies less than 1 week); however, if LMWHs would accumulate, this would be noticeable already after three dosages. Given the fact that patients tend to be discharged from the hospital within a few days, it is hardly possible to perform clinical studies with longer follow-up.

Another limitation might be that the risk of bias could not be assessed accurately. Considering bias across studies, we feel that a publication bias seems unlikely since both negative and positive outcomes in studies on accumulation have news value, and we included two abstracts of unpublished data. Furthermore, a possible confounder could be the difference in renal insufficiency onset (acute vs chronic) in different studies; however, there is no evidence that this can cause variance in anti-Xa activity.

A major limitation of all studies is that anti-Xa activity rather than hard clinical endpoints were studied. Although a study on hard clinical endpoints in patients with renal insufficiency is probably not feasible due to large numbers needed, it should be taken into account that the correlation between anti-Xa activity and occurrence of bleeding or VTE is not unambiguous [56–60]. The therapeutic and prophylactic target levels of anti-Xa activity are not supported by evidence of trials, but they are rather based on expert opinions [61, 62]. We also found in our review that in six studies that reported clinical outcomes, none of the patients with bleeding had higher anti-Xa activity than patients without bleeding [31–36]. However, the anti-Xa activity is considered to be the best test available to measure LMWH activity and to detect accumulation of LMWHs [4, 63].

In conclusion, for several LMWHs the guidelines that recommend dose reduction for prophylactic use in patients with renal insufficiency are evidence based, except for dalteparin, tinzaparin, and nadroparin. We recommend a dose reduction for prophylactic use of enoxaparin, bemiparin, and certoparin in patients with CrCl below 30 ml/min [24, 28, 29, 38, 39]. Prophylactic dosages of tinzaparin and dalteparin are likely to be safe in patients with renal insufficiency and do not need dose reduction. Studies are needed to assess accumulation of prophylactic dosages of nadroparin and for all LMWHs in patients undergoing renal replacement therapy.