Patient-controlled Analgesia and Postoperative Pressure Ulcer: A Meta-analysis of Observational Studies

Background. Patient-controlled analgesia (PCA) has become a common practice after surgery, but research has shown that the use of PCA is also a significant risk factor for pressure ulcers. However, no meta-analysis or conclusive review has investigated whether patients using PCA have a significantly higher prevalence of pressure ulcers. Objective. This study explores the association between the use of postoperative PCA and the prevalence of pressure ulcers. Materials and Methods. PubMed, the Cochrane Controlled Register of Trials, Web of Science, China National Knowledge Infrastructure, Wanfang, and Vip databases were searched to identify studies, published up until November 2016, concerning the association between PCA and pressure ulcer prevalence. A manual search of the references of relevant studies also was performed. Odds ratio (OR) and corresponding 95% confidence interval (CI) were used to evaluate the strength of association between the use of PCA after surgery and pressure ulcer prevalence. The methodological quality of included case-control studies and cohort studies was assessed by the Newcastle-Ottawa Scale. The test of heterogeneity, subgroup analysis, meta-regression, Begg’s funnel plot, and Egger’s test also were used. Results. Four cohort studies and 1 case-control study were included. In these 5 studies, 265 participants were identified. In pooled analysis, heterogeneity was 0 among the studies. In a fixed effects model, postoperative pressure ulcer was associated with PCA (pooled OR, 3.525; 95% CI, 1.655–7.509). Subgroup analysis of these 5 studies yielded an OR of 3.29 (95% CI, 1.47–7.40) for cesarean section, 5.10 (95% CI, 0.24–107.55) for general surgery, and 5.10 (95% CI, 0.24-107.55) for orthopedic surgery. There was no heterogeneity among the 5 studies. Additional meta-regression of year and incidence did not find significant outcomes. Conclusions. This meta-analysis shows PCA may be associated with an increased risk of postoperative pressure ulcer, especially after caesarean section. More evidence-based studies on this research field are needed to draw a firmer conclusion.

RELATED CONTENT

Stage IV Perineal Pressure Ulcers in Immobile Patients Treated With Surgical Flap Closure Augmented With Cryopreserved Placental Membrane Containing Viable Cells

Sacral Pressure Ulcer-induced Fournier’s Gangrene Extending to the Retroperitoneum: A Case Report

SCALE Wounds: Unavoidable Pressure Injury

Pressure ulcers are common in nursing care and lead to large economic pressure and societal health burdens.¹ It also has become a common complication from long-term bed rest following surgery, which may sharpen postoperative pain, increase risk of infection, and lengthen the duration of hospitalization.² Reports have shown a high risk of pressure ulcer incidence among surgical patients,³ as the incidence of surgery-related pressure ulcers ranged from 0.003 to 0.574.⁴

Risk factors associated with the etiopathogenesis of pressure ulcers, which can appear during surgery, can be classified into intrinsic and extrinsic factors. Extrinsic factors include surgery type and duration, anesthesia, and surgical positioning.³ Postoperative patient-controlled analgesia (PCA) has become a common practice after surgery. Nevertheless, research has shown the use of PCA is also a significant risk factor for pressure ulcers.⁵ Relative observational studies and case reports have shown pressure ulcers related to the use of postoperative PCA may occur even in young patients, such as puerpera after cesarean section, suggesting more preventive measures against pressure ulcers are necessary in puerpera using PCA.^6-12 However, whether the use of postoperative PCA will increase the incidence of pressure ulcers remains unclear, and the knowledge of the clinical importance and implications of these associations are limited.

As of now, there are few systematic reviews and meta-analyses reporting on the aforementioned associations. In order to evaluate the validity of the association between PCA and the risk of postoperative pressure ulcers, the authors conducted a meta-analysis of observational studies.

This article was reported according to the guidelines for Meta-Analyses and Systematic Review of Observational Studies (MOOSE).¹³

Literature search
The authors identified all relevant articles by searching PubMed, the Cochrane Controlled Register of Trials, Web of Science, China National Knowledge Infrastructure, WanFang, and VIP databases. The databases were searched for articles published from database inception up until November 23, 2016. Search strategies were tailored to each database. For example, Medical Subject Headings (MeSH) were used in combination with a free term to seek out more relevant studies when searching on PubMed; the following search algorithm was used: (“analgesia” [MeSH] OR “analgesia” [title/abstract]) AND (“pressure ulcer”[MeSH] OR “pressure ulcers”[title/abstract] OR “bedsore”[title/abstract] OR “pressure sore”[title/abstract] OR “decubitus ulcer”[title/abstract]). The authors also manually searched the references of all relevant studies to supplement the database searches. Only articles published in either English or Chinese were searched.

Eligibility criteria
To be included, prospective and retrospective cohort studies had to separately mention the condition and incidence of postoperative pressure ulcers on patients who used postoperative analgesia pump and patients who did not use postoperative analgesia pump. There was no limitation of publication time or pressure ulcer stage. Since the primary aim was to estimate the validity of the association between PCA and the risk of postoperative pressure ulcers, included studies must have consisted of at least 2 groups (group with PCA and group without PCA). Case reports, animal studies, and in vitro studies were excluded from the results.

Study selection and data collection process
Two researchers screened the titles and abstracts of relevant studies found on the databases according to the inclusion criteria; then, the researchers screened the full texts. A standardized data extraction form was created to collect key information from the relevant studies, including first author, year of publication, type of included study, number of included participants in each study, the kind of surgeries, analgesic tube, location of pressure ulcer, incidence of pressure ulcer of experimental group and control group, average age, and gender of participants. The same 2 reviewers independently extracted information from each article. Any differences were discussed and resolved by consensus. When needed, the reviewers contacted the original author for clarification. 

Quality assessment
The Newcastle-Ottawa Scale (NOS),¹⁴ a validated scale for nonrandomized studies in meta-analyses, was used to assess the quality of the included studies. The assessment scale allowed a total score of up to 9 points and consisted of 3 domains: selection of study groups (4 items: representativeness of the exposed cohort, selection of the nonexposed cohort, ascertainment of exposure, and demonstration that outcome of interest was not present at start of study), comparability (1 item: comparability of cohorts on the basis of the design or analysis), and outcome (3 items: assessment of outcome, follow-up length, adequacy of follow-up). In this scale, each term is followed by several choices, and a choice marked with a star represents good quality. One star represents one point; all items can be awarded a maximum of 1 point with the exception of comparability, which can receive 2 points. The low, moderate, and high quality of studies were graduated by scores of 0-3, 4-6, and 7-9.¹⁵ 

Statistical analysis 
A fixed effects meta-analysis weighted by the Mantel-Haenszel method was used to estimate pooled odds ratios (ORs) and 95% confidence intervals (CIs). Odds ratio was chosen as effect measures because the type of studies consisted of prospective and retrospective studies. Original data were converted to ORs. The heterogeneity was estimated with a Cochran Q test and an I² test. Subgroup analyses were completed for the type of surgery, and forest plots were used to evaluate overall effects. A meta-regression assessed the predictive effect of the variables on heterogeneity. The authors assessed publication bias by visual inspection of asymmetry in a funnel plot and an Egger’s test. All analyses were done with Stata 12.0 (StataCorp LLC, College Station, TX). 

Patient involvement
No patients were involved in the design of this meta-analysis; no ethical approval was required.

Study selection
Figure 1 shows the selection process of the articles included in the meta-analysis. The original search yielded 265 records; after removal of duplicates, 152 were reviewed. Following title/abstract screening, 42 articles remained for full-text review, and after a full-text review, 5 Chinese-language articles were included.

Study characteristics
Of all the included studies, 4 were cohort studies^6-8,16 and 1 was a case-control study.⁵ All studies used a comparison group and achieved a total NOS grading score (Table 1^5-8,16). As shown in Table 1,^5-8,16 2 studies^7,8 received 8 points; these studies lost 1 point because they did not state whether they used a blind study method. Two studies^6,16 received 6 points, losing 3 points for not stating the use of a blind study method and for not controlling for some confounding factors. The study conducted by Chen and Huang⁵ received 5 points; it lost 4 points because participants in the control group came from the hospital rather than the community and it also did not state whether they used a blind method. In addition, the representativeness of the cases in this study⁵ had potential for selection bias, and this study had no description of the control group.

Table 2^5-8,16 shows the overall characteristics of the included studies,^5-8,16 comprising 758 participants. Among these 5 studies, 3 focused on cesarean sections,^5,7,8 1 on general surgery,¹⁶ and 1 on orthopedic surgery.⁶ In the 4 prospective studies,^6-8,16 the number of pressure ulcers and participants in each group with PCA were 2/102, 1/50, 2/80, and 2/102, while the group without PCA were 0/102, 0/50, 0/80, and 0/102, respectively. In addition, the retrospective study⁵ had pressure ulcer-participant ratios of 23/40 in the group with PCA and 15/50 in the group without PCA. The studies were published between 2005 and 2011. 

Synthesis of results
Table 2^5-8,16 describes the primary outcomes for each study. Results of the individual studies were contained in a forest plot (Figure 2^5-8,16). There was no heterogeneity between the included studies (overall I² = 0.0%, P = .993), so a fixed effects model was used for this meta-analysis. The summary OR of patients with PCA compared with patients without PCA was 3.53 (95% CI, 1.65–7.51; Z = 3.27, P = .001)

Sensitivity analysis for the association between PCA after surgery and pressure ulcer prevalence was carried out to evaluate the reliability of this meta-analysis (Figure 3^5-8,16). The pooled ORs of sensitivity analysis, with each study excluded once individually, were 3.42 (95% CI, 1.56–7.46), 3.56 (95% CI, 1.63–7.74), 3.42 (95% CI, 1.56–7.47), 3.42 (95% CI, 1.56–7.47), and 4.60 (95% CI, 0.99–21.45).

Visual inspection of Begg’s funnel plot of the included studies and Egger’s test  were not suggestive of publication bias (P = .100; 95% CI, -0.111–0.736) (Figure 4).

Additional analyses
Figure 5^5-8,16 shows the subgroup analysis by type of surgery (cesarean vs. general vs. orthopedic). In the 3 studies on cesarean surgery,^5,7,8 the summary OR was 3.294 (95% CI, 1.466–7.402; Z = 2.89, P = .004). In the 1 study on general surgery,¹⁶ the summary OR was 5.100 (95% CI, 0.242–107.549; Z = 1.05; P = .295). In the 1 study on orthopedic surgery,⁶ the summary OR also was 5.100 (95% CI, 0.242–107.549; Z = 1.05; P = .295). The last 2 subgroups showed no statistical significance.

Meta-regression by publication year showed no significance (t = -0.07; P = .958), and meta-regression by pressure ulcer incidence also did not find significant outcomes (t = -0.08; P = .949) (Figures 6, 7).

This meta-analysis found that there may be an association between PCA and an increased incidence of pressure ulcers after surgery (pooled OR, 3.53; 95% CI, 1.66–7.51), especially after cesarean section (OR, 3.29; 95% CI, 1.47–7.40).

There are multiple possible reasons behind the correlation between PCA use and increased pressure ulcer incidence. First, the use of a PCA pump can relieve the sense of pain; a decreased sensation and blocked motor nerve owing to a PCA pump might accelerate the development of pressure ulcers.¹⁰ Meanwhile, movement actually can be considered self-protection for bedridden patients; however, PCA can reduce the patient’s sense of pain so that they will reduce their movement, which can reduce the patient’s ability to self protect. Similarly, the use of PCA may make patients overemphasize the disease severity so that they may fear moving. Second, because a PCA pump can relieve painful sensations, patients may be less willing to move in order to stay comfortable. Third, in order to avoid the events of catheter prolapse and compression, postoperative patients are fearful of moving.⁵

Results in relation to other studies
To date, few relevant meta-analyses of the association between PCA and postoperative pressure ulcers have been reported. Only a few previous prospective cohort studies,^6-8,16 case-control studies,^5,17 and case reports^9-11 have reported on both PCA and postoperative pressure ulcers. All of these studies^9-11 mentioned that the occurrence of pressure ulcers was associated with the use of PCA. Some studies^5,17 investigating the risk for or causes of pressure ulcers have verified this association, which was in line with the present results. For instance, Chen and Huang⁵ found the use of anesthesia pump was a high-risk factor associated with postoperative pressure ulcers (OR, 3.063; P < .05) by logistic regression analysis. Furthermore, Zhang et al¹⁷ investigated the causes and nursing prevention for pressure ulcers post cesarean section. They also concluded that the postoperative use of an analgesia pump is a risk factor for pressure ulcers and that the incidence of pressure ulcers in patients using an analgesia pump is 4-times that of patients without one.¹⁷

The present findings are generally consistent with the results of previous research^5,17 except for 1 case-control study.¹⁸ This case-control study¹⁸ investigated the risk factors of pressure ulcers for patients in prone position. It concluded that there was no statistical significance in the variation of pressure ulcer incidence between patients with and without PCA (P > .05). An unreasonable ratio of participants in the case group (patients with analgesia pump, n = 182) to the control group (patients without analgesia pump, n = 18) may account for this inverse result.¹⁸ Another reason for those results may be that this case-control study investigated pressure ulcers of patients in prone position while other studies^5-8,16 researched pressure ulcers of patients in supine position. Further research is needed to explore whether a patient's position following surgery is a risk factor of pressure ulcers. 

Implication for clinical practice
The current evidence suggests a higher risk of pressure ulcers in postoperative patients with PCA compared with patients without PCA. With respect to clinical care, the present results indicate that medical workers should pay closer attention to postoperative patients with an analgesia pump to prevent the occurrence and development of a pressure ulcer. The use of a risk assessment scale (RAS) for pressure ulcers has become a key point for prevention.¹⁹ The most commonly used RASs for pressure ulcer include the Norton scale, the Waterlow scale, the Braden scale, and the Braden revision scale.²⁰ However, most of these RASs usually are applied to older or chronic patients who need long-term bedrest rather than postoperative patients, because the RASs lack key influence factors for surgical operations.²¹ Thus, some new risk factors for pressure ulcers should be considered to establish a new RAS specifically for postoperative patients.

According to the present study, the use of PCA after surgery may be a new risk factor that can supplement existing pressure ulcer risk tools for inpatients, similarly to patient-reported symptoms, older age, and low albumin.²² This possible new risk factor is consistent with the recommendation of international guidelines from the National Pressure Ulcer Advisory Panel, the European Pressure Ulcer Advisory Panel, and the Pan Pacific Pressure Injury Alliance that state a RAS should not be used alone but rather in conjunction with consideration of other relevant risk factors.

In the current study, the quality of the included studies was not good. The majority^5,6,16 of these studies were only regarded as of moderate quality according to NOS, and the number of included studies was small due to a lack of relevant research on this topic, which may cause bias. It is possible that regression analysis did not find some significant outcomes due to the previously mentioned reasons. Also, in the subgroup analysis sorted by type of surgery (cesarean, general, orthopedic), only the results of cesarean participants found significant outcomes, which may result from the sole study included in the other 2 subgroups. Thus, more evidence-based studies are need.

However, there are several strengths to the present study. First, there are few relevant meta-analyses on the association between PCA and postoperative pressure ulcers. The present authors systematically identified relevant case-control and cohort studies to extract data on the association between PCA and postoperative pressure ulcers. Second, the authors used NOS to assess the quality of the included studies. Third, heterogeneity among the included studies was 0 and visual inspections of a Begg’s funnel plot and an Egger’s test were not suggestive of publication bias.

The present study found that PCA may be associated with postoperative pressure ulcers, especially after caesarean section. The majority of studies included in this meta-analysis had samples comprised of young women, though the number of included studies was small due to the lack of relevant original studies. Future general population research should investigate the association between PCA and postoperative pressure ulcers. More prospective studies in this research field are needed and meta-analyses of this issue should include more high-quality studies to draw firmer conclusions. 

Authors: Man-Li Zha, BS¹; Bao-Fang Yuan, BS²; Ji-Yu Cai, BS¹; Yi-Ping Song, BS¹; and Hong-Lin Chen, MD¹

Affiliations: ¹School of Nursing, Nantong University, Nantong, Jiangsu, China; and ²Affiliated Hospital of Nantong University, Nantong, Jiangsu, China

Correspondence: Hong-Lin Chen, MD, Dean, School of Nursing, Nantong University, QiXiu Road 19#, Nantong, Jiangsu Province 226001 China; honglinyjs@126.com

Disclosure: This work was supported by Nantong Municipal Science and Technology Bureau (grant MS 1201706-3). The authors disclose no conflicts of interest.