Background

IPF is commonly associated with fibrosing interstitial pneumonia and a poor prognosis [1,2,3]. Among adults over 65 years of age, Schafer SC et al. [4] reported 494 cases per 100,000 people, which is double the rate observed a decade ago. Patients diagnosed with IPF typically experience a median survival time of 2–5 years, and quite a few of them are not in the early stage when they see a doctor, resulting in a marked deterioration in their quality of life [5, 6]. However, drug treatment for IPF is not effective enough, and there is a pressing need for additional treatment modalities [1,2,3,4]. Pirfenidone has been shown to regulate transforming growth factors and fibroblasts, thereby producing antioxidant and antifibrotic effects, as well as regulating reactive oxygen species metabolism [7, 8]. Multiple RCTs have provided evidence that pirfenidone can effectively decelerate the progression of lung function and extend progression-free survival [9, 10]. NAC, an antioxidant and scavenger of oxygen free radicals, serves as a precursor of glutathione and can collaborates with catalase to decompose hydrogen peroxide into oxygen and water [11]. The addition of NAC to azathioprine and prednisone has been demonstrated to postpone the decline of lung function in the IFIGENIA study [11, 12].

Several case–control and cohort studies conducted in Japan, Germany, and China have demonstrated the efficacy of NAC plus pirfenidone for the treatment of IPF [13,14,15,16,17]. However, a number of high-quality RCTs, such as those conducted by Sakamoto S et al. and PANORAMA research, have revealed that the combination therapy does not offer superior efficacy compared to pirfenidone monotherapy for IPF treatment. Furthermore, the combination therapy has been found to increase the incidence of adverse effects [18, 19].

To mitigate the impact of confounding variables, and building upon the robust findings of Sakamoto S et al. [18], this study exclusively undertook a meta-analysis of RCTs pertaining to the use of NAC plus pirfenidone versus pirfenidone monotherapy for the management of IPF.

Methods

Data resources and literature search

The present systematic review and meta-analysis has been duly registered in PROSPERO (http://www.crd.york.ac.uk/PROSPERO), with the registration number CRD42023417130. Furthermore, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and PRISMA 2020-checklist were employed to carry out this systematic review and meta-analysis (Tables S1 and S2) [20]. A highly experienced librarian was enlisted to carry out the search strategy, which spanned from the inception of the database to April 5, 2023, and encompassed EMBASE, Cochrane Library, PubMed, Web of Science, and various Chinese databases (namely, Wanfang Database, CNKI, and VIP database). Additionally, searches were conducted on The EU Clinical Trials Register and clinicaltrials.gov to identify potential studies. The search terms employed were “N-acetylcysteine”, “acetylcysteine”, “pirfenidone”, “pulmonary fibrosis”, “idiopathic pulmonary fibrosis”, and “IPF” for all studies. For example, our search terms in Pubmed are:((N-acetylcysteine[All Fields]) OR (acetylcysteine[All Fields])) AND (pirfenidone[All Fields]) AND ((pulmonary fibrosis[All Fields]) OR (idiopathic pulmonary fibrosis[All Fields]) OR (IPF[All Fields])).We strictly follow the above search keywords to search, and there are no other restrictions in the database search.

Research selection

Inclusion criteria: The present study incorporated RCTs that evaluated the efficacy of pirfenidone plus NAC or pirfenidone monotherapy for the management of IPF. Exclusion criteria: Case–control studies, cohort studies, case reports, and other types of studies must be excluded. Additionally, we conducted comprehensive searches across pertinent networks but studies that do not meet the criteria must be excluded. Two independent researchers conducted title and abstract screening, and articles that met the eligibility criteria were subjected to full-text screening and ultimately included in the analysis.

Data extraction and quality assessment

Two researchers (XLZ and BZ) extracted the data independently. In the event of discrepancies, a third researcher is tasked with identifying and resolving them. When necessary, we consult pertinent authors and obtain relevant data. Any studies deemed unsuitable or incomplete are excluded from analysis. The parameters were extracted from qualifying studies: (1) Patient demographics, time of diagnosis, drug utilization, dosage, and duration; (2) Primary outcomes pertaining to changes in lung function (ΔFVC, ΔFVC%, Δ6MWT, ΔDLco%); and (3) Secondary outcomes: Incidence rates of at least one side effects, severe side effects, gastrointestinal effects, skin side effects, and mortality.

Risks of bias

Two researchers conducted independent assessments of the risk of bias in RCTs using the Cochrane Collaboration risk of bias tool [21]. The risk of bias was assessed across various domains, including random sequence generation, allocation concealment, blinding of participants and personnel, incomplete outcome data, selective reporting, and other biases. The ratings were categorized as either (I) low risk of bias, (II) probably low risk of bias, (III) probably high risk of bias, or (IV) high risk of bias. In the event of any discrepancies, a third researcher was consulted to facilitate resolution. The Cochrane Collaboration risk of bias tool was employed to assess five RCTs, revealing that 4 of them were deemed high risk, with the exception of the article authored by Behr J et al. (Table 3) [18, 19, 22,23,24].

Evidence quality GRADE evaluation

Two researchers independently and repeatedly assessed the certainty of the evidence using the GRADE approach (Table 1) [25, 26]. The evidence quality was assessed across various domains, including risk of bias, inconsistency, indirectness, imprecision, and publication bias. The ratings were categorized as either (I)extremely low certainty, (II) low certainty, (III) moderate certainty, or (IV) high certainty [27].

Table 1 GRADE quality assessment criteria
Full size table

Statistical analysis

The statistical analysis was conducted utilizing RevMan software (Version 5.0.1) [28]. The primary outcomes of this analysis encompassed ΔFVC, Δ%FVC, Δ6MWT, Δ%DLco, incidence of at least one side effects, incidence of severe side effects, incidence of gastrointestinal effects, incidence of skin side effects, and mortality from all causes in the treatment of IPF with pirfenidone plus NAC versus pirfenidone monotherapy. Mean and standard deviation were employed to express all data, while frequency and percentage were used to summarize the classification outcomes. Binary results were expressed using relative risk (RR) and 95% confidence intervals. Results of PFT parameters and other continuous variables were presented using Standardized mean difference (SMD). The level of heterogeneity was tested using the I2 statistic, with values exceeding 25%, 50%, and 75% indicating low, moderate, and high heterogeneity, respectively. All models apply random effects due to the inherent differences in populations. Unless otherwise specified, P < 0.05 is considered statistically significant [29, 30].

Results

Literature search and risk score

A total of 1246 relevant studies were searched and identified. Through the examination of titles and abstracts, 71 highly relevant articles were found. Following a meticulous full-text screening of the remaining 71 articles, 5 RCT articles were ultimately identified (Fig. 1) [18, 19, 22,23,24]. Notably, these 5 articles were conducted in various regions, including 1 in Japan, 1 in Europe, and 3 in China. 3 articles from China were published in Chinese, while the remaining articles were published in English. 2 were sourced from the EMBASE database [18, 19], and 3 were obtained from the CNKI database [22,23,24]. The analysis encompassed a total of 398 individuals, with 196 in the treatment group and 202 in the control group (Table 2).

Fig. 1
figure 1

Flowchart of the article evaluation process in this meta-analysis

Full size image
Table 2 Basic information of 5 RCTs
Full size table

Effect of NAC plus pirfenidone on outcomes of IPF

Five articles reported ΔFVC(NAC + PFD:n = 187, PFD alone:n = 195); We found that NAS plus pirfenidone may not improve ΔFVC as compared to pirfenidone monotherapy for IPF with high heterogeneity(SMD 0.18; 95%CI -0.68 to 1.05, P = 0.68, I2 = 94%; extremely low certainty) (Fig. 2 and Table 4) [18, 19, 22,23,24]. Three articles reported Δ%FVC(NAC + PFD:n = 106, PFD alone:n = 112); We found that NAS plus pirfenidone may not improve Δ%FVC as compared to pirfenidone monotherapy for IPF with high heterogeneity(SMD -2.62, 95%CI -5.82 to 0.59, P = 0.11, I2 = 99%; extremely low certainty) (Fig. 3 and Table 4) [18, 19, 24]. Three articles reported Δ6MWT(NAC + PFD:n = 107, PFD alone:n = 112); We found that NAS plus pirfenidone may not improve Δ6MWT as compared to pirfenidone monotherapy for IPF with high heterogeneity(SMD -0.35, 95% CI -0.98 to 0.28, P = 0.28, I2 = 80%; extremely low certainty) (Fig. 4 and Table 4) [18, 19, 24]. Four articles reported Δ%DLco(NAC + PFD:n = 143, PFD alone:n = 154); We found that NAS plus pirfenidone may not improve Δ%DLco as compared to pirfenidone monotherapy for IPF with moderate heterogeneity(SMD -0.17, 95% CI -0.15 to 0.48, P = 0.29, I2 = 45%; low certainty) (Fig. 5 and Table 4) [18, 19, 22, 24].

Fig. 2
figure 2

ΔFVC forest plot

Full size image
Fig. 3
figure 3

Δ%FVC forest plot

Full size image
Fig. 4
figure 4

Δ6MWT forest plot

Full size image
Fig. 5
figure 5

Δ%DLco forest plot

Full size image

Safety profile of NAC plus pirfenidone in the treatment of IPF

Four articles reported the incidence of at least one side effect(NAC + PFD:n = 153, PFD alone:n = 159); We found that NAS plus pirfenidone may not reduce the incidence of at least one side effects as compared to pirfenidone monotherapy for IPF with low heterogeneity(RR 1.00, 95%CI 0.84 to 1.19, P = 0.98, I2 = 0%; low certainty) (Fig. 6 and Table 4) [18, 19, 23, 24]. Five articles reported the incidence of severe side effects(NAC + PFD:n = 196, PFD alone:n = 200); We found that NAS plus pirfenidone may not reduce the incidence of severe side effects as compared to pirfenidone monotherapy for IPF with low heterogeneity(RR 0.67, 95%CI 0.30 to 1.47, P = 0.31, I2 = 0%; low certainty) (Fig. 7 and Table 4) [18, 19, 22,23,24]. 4 articles reported the incidence of gastrointestinal effects(NAC + PFD:n = 153, PFD alone:n = 159); We found that NAS plus pirfenidone may not reduce the incidence of gastrointestinal effects as compared to pirfenidone monotherapy for IPF with low heterogeneity(RR 0.67, 95%CI 0.41 to 1.09, P = 0.11, I2 = 0%; low certainty) (Fig. 8 and Table 4) [18, 19, 23, 24]. 4 articles reported the incidence of skin effects(NAC + PFD:n = 153, PFD alone:n = 159); We found that NAS plus pirfenidone may not reduce the incidence of skin effects as compared to pirfenidone monotherapy for IPF with low heterogeneity(RR 1.26, 95%CI 0.64 to 2.45, P = 0.50, I2 = 0%; moderate certainty) (Fig. 9 and Table 4) [18, 19, 23, 24]. Five articles reported the incidence of mortality(NAC + PFD:n = 196, PFD alone:n = 202); We found that NAS plus pirfenidone may not reduce the incidence of mortality as compared to pirfenidone monotherapy for IPF with low heterogeneity(RR 0.35,95%CI 0.07 to 1.68, P = 0.19, I2 = 0%; moderate certainty) (Fig. 10 and Table 4) [18, 19, 22,23,24].

Fig. 6
figure 6

Side effects forest plot

Full size image
Fig. 7
figure 7

Severe side effects forest plot

Full size image
Fig. 8
figure 8

Gastrointestinal effects forest plot

Full size image
Fig. 9
figure 9

Skin effects forest plot

Full size image
Fig. 10
figure 10

Mortality rates forest plot

Full size image

Bias and sensitivity analysis

If the funnel plot of the meta-analysis describes that most studies are located in the upper region of the “inverted funnel” and show a rough symmetry, it indicates that publication bias is not significant, and vice versa, publication bias is significant. According to the findings from the funnel plot analysis, no substantial bias was observed in the publication of ΔFVC, Δ6MWT, Δ%DLco, skin effects, and mortality rates. However, a significant bias was detected in the publication of Δ%FVC, side effects, severe side effects, and gastrointestinal effects. However, the number of articles is limited, which limits the significance of funnel plot to some extent. There were only 5 relevant RTC trials, 2 for inhalation administration and 3 for oral administration. In order to reduce heterogeneity and increase sensitivity, oral administration studies was performed for analysis, and the forest plots can be found in the Supplement Material (Figs. S10, S11, S12, S13, S14, S15 and S16). However, subgroup analysis failed to significantly reduce heterogeneity and change results.

Discussion

In this meta-analysis, by analyzing the results of ΔFVC, Δ%FVC, Δ6MWT, and Δ%DLco, NAC plus pirfenidone is unlikely to exhibit greater efficacy compared to pirfenidone monotherapy in the treatment of IPF. The initial proposal for NAC as a treatment for pulmonary fibrosis was based on findings from an animal model. NAC has been shown to enhance glutathione synthesis, regulate lysine oxidase activity, inhibit epithelial-mesenchymal transformation, and ultimately reduce the presence of oxygen free radicals and provide antioxidant effects [11, 31,32,33]. The efficacy of NAC monotherapy in the treatment of IPF has been controversial [34,35,36]. The meta-analysis conducted by Sun et al. [37] revealed that NAC had a significant impact on reducing vital capacity and Δ6MWT. However, it did not significantly reduce Δ%DLco, ΔFVC, adverse events, or mortality. Previous cohort studies and case controls have demonstrated that NAC combined with other medications has a certain efficacy in treating IPF [11,12,13,14,15,16,17]. In previous studies, 5 articles were cohort and case–control studies comparing the treatment of pirfenidone puls NAC versus pirfenidone monotherapy for IPF [13,14,15,16,17]. The conclusions reached by these articles are inconsistent. Shi H et al. [38] systematically reviewed these 5 articles and did not find that pirfenidone plus NAC had better benefits in the treatment of IPF. Nevertheless, a phase 3 RCT conducted by Sakamoto et al. [18] in Japan found that NAC plus pirfenidone did not differ from pirfenidone monotherapy in terms of Δ%DLco, Δ6MWD, and progression-free survival in IPF treatment, and ΔFVC decreased more significantly. A Phase 2 RCT in Europe by Behr J et al. [19] showed that oral NAC plus pirfenidone was unlikely to have clinical benefit. Related RCTs have also been conducted in China, but with relatively high risk of bias (Table 3). Wen JY et al. and Zhou XD et al. showed that NAC plus pirfenidone was effective in the treatment of IPF [22, 23]. However, the results of Zhao HM et al. showed that it had no better efficacy [24]. But a common shortcoming of these studies is the small number of patients. According to the PANTHER study [35], NAC may be effective in IPF treatment. However, subsequent analysis revealed its benefits primarily for patients with T/T genotypes, while patients with C/C genotypes may not experience the same benefits and could potentially experience disease progression. The study speculates that genes encoding T-cell interaction proteins(TOLLIP; Rs3750920) may influence the response to NAC [35, 36]. This may explain why some studies have shown that NAC is effective in treating IPF, but more relevant studies are needed to confirm this.

Table 3 RCT risk assessment using the Cochrane collaboration risk of bias tool
Full size table

This meta-analysis showed that NAC plus pirfenidone was not more effective than pirfenidone monotherapy in the treatment of IPF in terms of at least one side effect, severe side effects, gastrointestinal effects, skin effects, and mortality rates. The side effects of NAC plus pirfenidone in the treatment of IPF have often been reported in different studies. The prevailing view is that the deaths are not related to medical treatment and are mostly the result of an exacerbation of the disease or other causes. Furthermore, severe side effects were predominantly unrelated to the administered treatment. Certain gastrointestinal effects were found to be related to the treatment, while others were not. However, these effects were generally mild to moderate in severity and could be alleviated through symptomatic treatment. The PANORAMA study indicated that the treatment group exhibited a higher incidence of photosensitivity compared to the control group, which was hypothesized to be associated with NAC [19]. Nevertheless, this meta-analysis did not reveal any significant difference in this regard [18, 19, 22,23,24]. These findings suggest that NAC plus pirfenidone in the treatment of IPF is relatively safe, although further confirmation is required through large-scale studies.

GRADE evidence grade evaluation results showed that moderate certainty evidence indicates that NAS plus pirfenidone as compared to pirfenidone monotherapy for IPF may not reduce the incidence of skin effects and mortality. Low certainty evidence indicates that NAS plus pirfenidone as compared to pirfenidone monotherapy for IPF may not improve Δ%DLco, ΔFVC, Δ%FVC, and Δ6MWT. Extremely low certainty indicates that NAS plus pirfenidone as compared to pirfenidone monotherapy for IPF may not improve ΔFVC, Δ%FVC, and Δ6MWT (Table 4) [25, 26].

Table 4 GRADE evidence quality evaluation of outcomes
Full size table

This study also exhibits evident limitations, including small sample sizes, limited geographical coverage, low statistical power and publication bias, despite being RCTs. Moreover, there are no high certainty outcomes in this paper; variations in the administration of NAC were observed across different studies, with two studies employing inhaled NAC, leading to higher concentrations of NAC in the pulmonary system. Furthermore, the data pertaining to study outcomes is incomplete, with certain studies lacking partial PFT data and some studies lacking information on side effects. These observations underscore the necessity for high-quality, large-sample RCTs to furnish more robust clinical recommendations, but this is also very difficult.

Conclusion

There is limited evidence that NAC plus pirfenidone is not more beneficial than pirfenidone monotherapy in the treatment of IPF in terms of ΔFVC, Δ%FVC, Δ6MWT, Δ%DLco, at least one side effect, severe side effects, gastrointestinal effects, skin effects, and mortality rates. Consequently, until more robust evidence becomes available, it is not advisable to routinely administer NAC plus pirfenidone for IPF. However, due to low statistical power and the existence of heterogeneity, objective explanations are also needed.