Over one-third of Th/To antibody positive scleroderma patients develop pulmonary hypertension in long-term follow-up

Abstract

Objective:

We describe the long-term follow-up and clinical manifestations of Th/To positive systemic sclerosis (SSc) patients.

Methods:

We performed a case-control study. Cases were new anti-Th/To antibody positive referrals to Pittsburgh Scleroderma Center from 1980–2015. Controls were the next two consecutive anti-Th/To negative patients seen after a case. Long-term disease manifestations and survival were compared.

Results:

204 anti-Th/To positive SSc patients were identified and 408 controls. The mean age of the entire cohort was 52 (±12.9) years and 76% were female. Anti-Th/To positive patients more often presented without skin thickening (p<0.0001) and had higher rate of pulmonary hypertension (PH) (p<0.0001) and ILD (p=0.05) than controls. They had less frequent muscle and joint involvement (p<0.0001). After a median clinical follow-up of 6.1 years (IQR: 2.4, 12.7), 38% of Th/To patients had developed PH compared to 15% of other SSc patients (p<0.0001). The rates of WHO Group 1 pulmonary arterial hypertension (PAH) were 23% in Th/To patients versus 9% in controls (p<0.0001). After adjustment for age and gender, anti-Th/To antibody was associated with a 3.3 (95% CI 2.3–4.9) increased risk of developing PH at 10 years of follow-up from the first SSc center visit.

Conclusion:

This is the largest cohort of anti-Th/To positive SSc patients with long-term follow-up data. Striking is the very high rate (38%) and associated independent risk of anti-Th/To patients developing PH in follow-up, particularly Group 1 PAH. Patients presenting with limited skin involvement should be tested for Th/To antibody. If present, careful monitoring for PH is warranted.

INTRODUCTION

Patients with systemic sclerosis (SSc) can be classified into two primary clinical subsets based on the extent of skin involvement: diffuse cutaneous and limited cutaneous. The spectrum of limited SSc includes patients who have no skin thickening, termed scleroderma sine scleroderma (ssSSc). There are SSc-specific serum autoantibodies with recognized associations to SSc clinical features (i.e. cutaneous subtype, internal organ involvement), and prognosis in SSc [1–5]. When both cutaneous subset and autoantibody status are known, a clinical-serologic classification can be used to inform the natural history of disease in groups of SSc patients.

Anti-Th/To is an uncommon SSc-associated anti-nuclear antibody (ANA) in a nucleolar pattern. The Th/To antigen consists of two RNA-processing enzymes (RNase MRP and RNase P) plus ten associated proteins, of which Rpp25, Rpp38, and hPop1 are the main autoantigens [6–9]. Recently, testing for the anti-Th/To antibody, which was historically accessible only to a few research centers, has become available on some commercial platforms, thereby increasing the need for clinical information on this subset of SSc patients.

The first series of SSc patients with anti-Th/To antibodies was reported by us in 1990 and included only 15 (4%) anti-Th/To positive patients out of 371 consecutive SSc patients between the years of 1984–1988 [10]. In 2002, we performed a follow-up study, comparing 107 SSc patients with anti-Th/To antibody to 365 SSc patients with anti-centromere antibody (ACA) seen from 1985–2000 [11]. Almost all anti-Th/To antibody-positive patients had limited cutaneous SSc and an increased frequency of interstitial lung disease (ILD) and pulmonary arterial hypertension (PAH) [11]. Other reports on Th/To positive SSc patients have consisted of small cross-sectional studies and are thus limited in their generalizability [6,9,12–23]. The long-term frequency of clinical outcomes and prognosis in SSc patients with anti-Th/To antibody are completely unknown.

The World Health Organization (WHO) currently classifies pulmonary hypertension (PH) into five groups based on etiology [24]. Systemic sclerosis patients can present with pulmonary arterial hypertension (PAH or Group 1), PH due to left heart disease (Group 2), or PH due to chronic lung disease such as ILD (Group 3). The relative frequency of these PH groups in patients with SSc and their correlations with SSc-associated serum autoantibodies are topics of current debate. However, patients with SSc are widely recognized to be at increased risk of pulmonary hypertension compared to the general population.

The objective of this study was to describe the long-term follow-up clinical features of SSc patients with anti-Th/To antibody, including the subclassification of PH into the WHO classification groups.

PATIENTS & METHODS

Patient Selection:

With informed consent, we included all patients who were anti-Th/To antibody-positive seen for an initial visit at the Pittsburgh Scleroderma Center between January 1, 1980 and December 31, 2015. There was no restriction in cutaneous subtype. We excluded patients with anti-Th/To antibody who also had another SSc-specific antibody to focus on the clinical associations of anti-Th/To. All patients included met the updated ACR/EULAR 2013 classification criteria over the course of their follow-up [25]. This study was reviewed and approved by the Institutional Review Board under approval number CR19090054–003.

Laboratory Methods:

Antinuclear antibody (ANA) testing was performed by indirect immunofluorescence (IIF). Anti-Th/To was detected by RNA immunoprecipitation (IP) as previously reported [11]. ACA was identified by its characteristic staining on HEp-2 substrate at 1:40 dilution. Anti-topoisomerase I and anti-U1RNP were tested for by Ouchterlony immunodiffusion as in previous studies [13]. All other SSc autoantibodies (anti-RNA polymerase III, anti-PM-Scl, anti-U3RNP, anti-U11/U12RNP and anti-Ku) were examined by protein immunoprecipitation or a combination of multiple methods as previously reported by Kao et al [26].

Study Design:

We used a 1:2 case-control design. To control for temporal trends, we matched each case to the next two consecutive SSc patients as controls, who could have any antibody profile excluding anti-Th/To. All cases and controls resided in the United States so that vital status during the follow-up period could be most easily determined.

Data Collection:

All patients had a comprehensive first visit evaluation collecting data on symptoms, physical examination, laboratory and serologic testing, and objective tests for internal organ dysfunction (pulmonary function tests [PFTs], chest CT scan, transthoracic echocardiogram (TTE), esophagram, etc). At follow-up, all patients completed an abbreviated 80-variable form. The electronic medical record system was retrospectively reviewed for objective testing and clinical outcomes not captured previously. We sent a follow-up questionnaire regarding SSc complications and current medications to cases and controls who had not been recently seen along with a request for outside medical records to supplement data. Outside medical records received due to this inquiry were reviewed for missing objective testing data (chest imaging, PFTs, TTE). Every effort was made to obtain objective test results to determine the frequency and severity of organ involvement.

Survival:

Survival as of December 31, 2019 was obtained from the Social Security Death Index. Cause of death was determined by review of medical records or discussion with managing physicians. SSc-related causes of death were categorized as pulmonary hypertension, pulmonary fibrosis, SSc-related kidney, heart, and kidney/heart disease combined. Non-SSc-related causes of death were categorized as cancer, sudden death, atherosclerotic heart disease, other non-SSc-related, unknown, infection, vasculitis, central nervous system disease. If records were incomplete or unavailable, the National Death Index (NDI) was used to obtain death certificates, which were reviewed in the context of known clinical information. Additional contact with medical personnel or family members, if available, was made to clarify cause of death.

Definitions of Organ Manifestations:

Organ system manifestations were defined as previously described [27]. ILD was defined as bibasilar fibrosis on chest radiograph or high-resolution CT scan of the lungs. PH was defined as mean pulmonary artery pressure >20 mmHg on right heart catheterization [24] or TTE with an estimated peak pulmonary artery systolic pressure >45 mmHg with determination of WHO subtype assessed in consultation with a PH expert.

Statistical Analysis:

Baseline characteristics and clinical features at baseline and follow-up were compared between the case and control groups using t-tests, Chi-square and non-parametric tests (including Fisher’s exact test and Mann Whitney Wilcoxon tests) where appropriate.

The risk of pulmonary hypertension and cumulative survival were assessed using time-to-event analysis by the Kaplan-Meier method. More specifically, we performed the following survival or time-to-event analyses to compare Th/To positive cases vs. controls: 1) 5-year survival from first visit; 2) 5-year survival from disease onset; 3) time-to-event analysis for all PH types from first visit over 10 years of follow-up; 4) time-to-event analysis for all PH types from disease onset over 20 years of follow-up; and 5) time-to-event analysis for WHO Group 1 PH from disease onset over 20 years of follow-up. Cox proportional hazard modeling was used to generate unadjusted and adjusted (for age and gender) hazard ratios.

All statistical analyses were performed using SAS 9.4 (Cary, North Carolina). Two-sided p-values ≤0.05 were considered statistically significant.

Patient and public involvement:

This study used a pre-existing and longstanding observational cohort over 30 years. Patients and the public were not involved in recruiting for or design and conduct of this study.

RESULTS

In total, we evaluated 3613 new SSc patients between 1980 and 2015. They included 211 SSc patients (5.6%) who were anti-Th/To antibody positive, of which 7 were excluded for having a concomitant second SSc-associated antibody, leaving 204 for analysis (Figure 1). We then identified 408 anti-Th/To negative controls.

Figure 1.

Patient Selection Flow Chart

Table 1 shows baseline demographics, length of follow-up, and cutaneous subtype of the anti-Th/To cases and controls. There were no significant differences in age, gender, or race between the groups. Cases were more likely to be prior or current smokers (64%) than controls (44%; p<0.0001). Nearly all (97%) anti-Th/To patients had limited skin thickening, with 23% presenting with ssSSc. Patients with the anti-Th/To antibody had significantly longer disease duration at presentation than those with other SSc antibodies (median 7.9 vs 3.3 years respectively; p<0.0001). Median length of follow-up of the entire cohort was 6.1 years (IQR: 2.4, 12.7), which did not differ between cases and controls. The most frequent first SSc-associated symptom was different between cases and controls (p<0.0001), as Th/To positive SSc patients overwhelmingly reported Raynaud phenomenon (80%) compared to 56% of controls. Other common first symptoms that differed included puffy fingers/hands in 6% of Th/To cases but 16% of controls and joint symptoms in 2% of Th/To patients versus 11% of controls.

Table 1.

Demographic and disease classification features of anti-Th/To positive (cases) and negative (controls) systemic sclerosis patients at first visit

	Cases (n=204)	Controls (n=408)	p-value
Demographic characteristics
Mean age (±SD) in years	52.6 (±12.0)	51.8 (±13.4)	NS
Gender (female)	79%	75%	NS
Race (Caucasian)	96%	91%	NS
Tobacco use			<0.0001
lifelong nonsmoker	36%	56%
prior smoker	37%	31%
current smoker	27%	13%
Disease classification
Diffuse cutaneous	3%	44%	<0.0001
Limited cutaneous	97%	56%	<0.0001
Sine scleroderma	23%	9%	<0.0001
Overlap syndrome	5%	11%	0.02
Clinical follow-up data
Median disease duration from onset to first visit (IQR) in years	7.9 (2.9, 14.9)	3.3 (1.2, 10.5)	<0.0001*
Median length of time from first visit to last (IQR) in years	5.5 (1.8, 12.8)	6.3 (2.7, 12.7)	0.20*

^*Mann-Whitney Wilcoxon test p-value is given for this continuous variable instead of the T-test p-value.

Organ system manifestations at baseline and as of the last follow-up are shown in Table 2, with pulmonary hypertension (PH) presented separately. At baseline joint and tendon manifestations were statistically less frequent in the Th/To positive patients, although ILD on imaging was more frequent (45%) than in control (other SSc antibody) patients (34%; p=0.05). No other significant differences in baseline organ involvement were present.

Table 2.

Organ system manifestations in anti-Th/To positive (cases) and negative (controls) SSc patients at baseline and last follow-up

	Baseline			Last follow-up
Organ system manifestations	Cases (n=204)	Controls (n=408)	p-value	Cases (n=204)	Controls (n=408)	p-value
Raynaud phenomenon	95% (n=192)	91%	NS	99%	97%	NS
Joint/tendon involvement	32% (n=64)	69% (n=275)	<0.0001	50%	80% (n=322)	<0.0001
Joint swelling	3% (n=6)	13% (n=50)	0.0001	12% (n=25)	25% (n=100)	0.0004
Tendon friction rubs	2% (n=4)	23% (n=91)	<0.0001*	3%	28% (n=112)	<0.0001*
Skeletal myopathy	1% (n=3)	5% (n=20)	0.04	2% (n=5)	6% (n=26)	0.04
Gastrointestinal involvement	38% (n=77)	46% (n=185)	NS	54% (n=109)	60% (n=246)	NS
Pulmonary fibrosis on imaging	45% (n=72)	34% (n=54)	0.05	54% (n=103)	39% (n=137)	0.0008
Renal crisis	2%	4%	NS*	3%	10%	0.003

^*Fisher’s exact test p-value is given for this categorical variable instead of the chi-square test p-value.

In follow-up, significantly more Th/To patients continued to have ILD on imaging (54%) vs controls (39%; p=0.008), with 93% of Th/To patients and 87% of controls having chest imaging available for review. Importantly, there was no difference in ILD severity using the Medsger severity scale [28] between Th/To positive patients and controls (p=0.42). Th/To patients continued to have statistically significant lower frequencies of joint and tendon manifestations, and over time had lower frequencies of renal crisis (3%) than the SSc control group (10%; p=0.003). Very few Th/To patients had muscle involvement at baseline (1%) or in follow-up (2%), compared to controls at baseline (5%; p=0.04) or follow-up (6%; p=0.04). Given the high rates of ILD on imaging, and the potential theoretical contribution of esophageal reflux to pulmonary radiographic changes, we compared GI severity using the revised Medsger severity scale between groups. Over time, there was a difference in GI severity between the groups (p<0.001) with only 7% having moderate, 4% severe and 2% end-stage GI manifestations in the Th/To group, whereas 12% had moderate, 3% severe and 5% end-stage in the control group.

Pulmonary hypertension frequency data are presented in Table 3. The few patients who had PH attributed to non-SSc related valvular disease (n=3) were not included in this analysis. Overall, SSc-related PH was detected at first SSc clinic visit more frequently in anti-Th/To antibody positive patients than in controls (25% vs. 9%, p<0.0001), increasing up to 38% during follow-up, compared to only 15% in controls (p<0.0001). Group 1 PAH was most common, with 17% of Th/To patients having Group 1 PAH, increasing up to 23% at last follow-up. In controls only 5% of patient had Group 1 PAH at baseline and 9% as of last follow-up (Table 3). Pulmonary hypertension risk: Due to the potential of survival bias given the longer disease duration of Th/To patients at baseline, we evaluated outcomes at 10 years of follow-up from the first visit. Kaplan-Meier analysis. This revealed an increased probability of development PH over 10 years of follow-up from first visit (p<0.0001, Figure 2) in Th/To patients.

Table 3.

Pulmonary Hypertension (PH) in Anti-Th/To Positive Patients and Controls at First Visit and after Follow-up

	Baseline			Follow-up
	Cases (n=204)	Controls (n=407)	p-value	Cases (n=204)	Controls (n=402)	p-value
No PH	74% (n=153)	91% (n=372)	<0.0001	62% (n=127)	85% (n=340)	<0.0001
Pulmonary arterial hypertension (Group 1)	17% (n=35)	5% (n=20)		23% (n=47)	9% (n=37)
PH related to cardiac disease (Group 2)	<1% (n=1)	<1% (n=2)		2% (n=4)	<1% (n=3)
PH related to lung disease (Group 3)	8% (n=15)	3% (n=13)		13% (n=26)	5% (n=22)

Figure 2.

Development of PH over 10 years of follow-up from first SSc Center visit. Number at risk is indicated at the bottom.

In Cox proportional hazard modeling, after age and gender adjustment, anti-Th/To antibody positivity was associated with a hazard ratio of 3.3 (95% CI 2.3–4.9) for the development of any PH. With modeling for the development of Group 1 PAH specifically, Th/To antibody positivity had 4.9 times higher (95% CI 2.6–9.4) risk with age. When additionally adjusted for disease duration, the risk of developing Group 1 PAH at 10 years from SSc-associated symptom onset remained quite similar at 5.1 (95% CI 2.7–9.8).

As anti-centromere antibody has been associated with a significant risk of developing group 1 PAH, we performed an additional Cox proportional hazards modeling for the development of PAH at 10 years including both anti-Th/To and ACA positivity. After adjustment for age, gender and disease duration, anti-Th/To positivity had a higher hazard ratio of 8.5 (95% CI 3.6–20.0; p<0.0001) for developing PAH, and ACA a lower hazard risk of 4.2 (95% CI 1.4–12.6; p=0.01).

Autoantibody profile:

Amongst controls 27% (n=110) were anti-centromere positive, 26% (n=106) anti-Scl70 positive, and 25% (n=101) RNA polymerase III positive. Anti-PM/Scl antibody was found in 7% (n=28), with anti-U1-RNP also found in 7% (n=28). Three percent (n=12) were anti-U3-RNP positive, 2% (n=9) anti-U11-RNP positive, 1% (n=6) anti-RUV-BL1/2 positive, and 1% (n=5) anti-Ku positive. Twenty-five patients had other antibodies (none Th/To, which did not exclude them from analysis) and one did not have serum available.

Survival:

As of last follow-up, 56% of Th/To patients and 55% of controls had died. Despite higher rates of PH at presentation, the 5-year cumulative survival from first Pittsburgh visit was 68% in anti-Th/To cases was significantly lower than controls at 76% (p=0.02) as shown in Figure 3.

Figure 3.

Five-year cumulative survival from first visit is lower in Th/To positive SSc patients than SSc (Th/To negative) controls. Number at risk is indicated at the bottom.

Cause of Death:

We were able to ascertain the causes of death in 101/113 (89%) of Th/To cases, and 164/224 (73%) of controls (Table 4). In both groups, most deaths were related to SSc: 61% of deaths in cases and 55% in controls were SSc-related. Overall, there was no difference in proportion of deaths attributed to SSc-related and non-SSc-related causes between the groups (p=0.3). The most frequent cause of death in anti-Th/To patients was PH (28%) and in controls was pulmonary fibrosis (18%).

Table 4.

Causes of death in anti-Th/To antibody–positive SSc cases and anti-Th/To antibody–negative controls^*

	Cases (n = 101)	Controls (n = 164)
SSc-related†	62 (61)	90 (55)
Non–SSc-related‡	39 (39)	74 (45)

^*Values are the number (%) of patients.

^†Systemic sclerosis (SSc)–related causes of death include pulmonary hypertension, fibrosis; SSc-related kidney disease, heart disease, and kidney/heart disease combined; SSc-related gastrointestinal disease; and other SSc-related causes of death.

^‡Non–SSc-related causes of death include cancer, sudden death, atherosclerotic heart disease, other non–SSc-related causes of death, unknown, infection, vasculitis, and central nervous system disease.

DISCUSSION

We report here on the clinical characteristics in long-term follow-up of 204 SSc patients with a positive anti-Th/To antibody, compared with 408 temporally matched SSc patients without this antibody. We have shown that patients with anti-Th/To antibody have a significantly higher risk of developing PH of any WHO group compared to other SSc patients in long-term follow-up. This is of seminal importance, as over one-third of all Th/To patients developed PH in long-term follow-up. These findings are relevant, as testing of the Th/To antibody using immunoprecipitation, as we have done here, is now commercially available.

In the recently published literature, several SSc studies have described the frequency of the anti-Th/To antibody in various populations. Incidence of the anti-Th/To antibody ranged from 2–8% in studies in American, Algerian, and Han Chinese SSc patients [29–31]. A study comparing a total of 260 African-Brazilian and white Brazilian SSc patients found a nonsignificant higher proportion of anti-Th/To positivity (5% compared to 2%) among white Brazilians [32]. Thus, anti-Th/To antibody occurs in a small proportion of SSc patients world-wide, except for its possible rarity or absence in sub-Saharan Africa [33] and the Hispanic populations of Central and South America (not yet studied). We found an overall 5% of patients with anti-Th/To positivity in 35 years of new patient visits in Pittsburgh.

Higher frequencies of ILD have been observed in patients with anti-Th/To antibody compared to controls in prior studies [11,12]. In these reports the controls used for comparison were ACA positive patients only, and the frequencies were 38% and 48% in anti-Th/To patients compared to 5% and 13% in ACA-positive patients, respectively. These results are expected as multiple reports in the literature have shown that lower rates of ILD in ACA-positive SSc patients [34,35]. We confirmed this finding in long-term follow-up of patients as the cumulative frequency of ILD (by imaging) in our population was 54% in anti-Th/To patients and 39% in controls. However, there was no difference in ILD severity on the Medsger severity scale between groups.

The findings of a nearly 5-fold increased risk of Th/To positive patients developing PAH within 10 years of their first SSc symptom are noteworthy. The increased risk is congruent with our prior shorter-term follow-up of anti-Th/To patients [11] which demonstrated higher frequencies of PAH in anti-Th/To positive patients. In that study, the comparator group was ACA-positive patients, although ACA positivity has been associated with PAH risk in the literature [36]. Higher frequencies of PAH or PH have not been reported in other case series of anti-Th/To positive patients, although the numbers of patients included in these studies were small [12,37]. Thus, our finding of a nearly 3-fold increased risk of Th/To positive patients developing PH of any WHO classification (Groups 1, 2 or 3) at 10 years of follow-up compared to other SSc patients is a new observation. What is of highest clinical relevance is the high rate of over one-third (37%) of Th/To positive patients developing some type of PH on long-term follow-up. As our mean clinical follow-up was only 6 years, it is probable that on longer-term follow-up this rate is even higher.

We also report lower rates of skeletal muscle and joint/tendon involvement in anti-Th/To positive patients compared to other SSc patients. Lower rates of esophageal involvement were also described in earlier papers [11,37]. In this study we found similar rates of GI involvement in both groups, although higher GI severity scores were less frequent in Th/To positive patients. Renal crisis did occur in 3% of anti-Th/To positive patients, but this was less than controls (10%) in long-term follow-up. This differs from the prior Pittsburgh series [11], perhaps related to the control group of ACA positive patients, whose risk of SRC is well known to be very low. Other small case series have not reported increased frequencies of renal crisis.

Of note, only 5 (2.5%) anti-Th/To positive patients developed diffuse skin involvement. Thus, the overwhelming majority of anti-Th/To antibody patients have limited or no skin involvement, consistent with other case series [12,37].

Finally, we reported a lower five-year cumulative survival from the first SSc visit in anti-Th/To positive patients compared to the controls which persisted after age and gender adjustment. Due to a significant difference in disease duration at first visit, survival from first SSc visit was analyzed from first visit rather than from symptom onset. This finding is in accordance with prior papers which suggest reduced survival in anti-Th/To positive patients [11]. The most frequent SSc-associated causes of death in the anti-Th/To positive and control patients were PH and ILD, in accordance with recent publications on cause of death in SSc [14,15]. Although not statistically significant, there were more PH-related causes of death in the anti-Th/To group compared to controls, as would be expected given the higher frequency of PH, regardless of WHO class.

The primary strength of this study is the large number of anti-Th/To positive patients with detailed long-term follow-up data, by far the largest case series reported to date. Another strength is the use of RNA immunoprecipitation, the gold standard method, to detect anti-Th/To antibody. One limitation is that the cases derive from a single tertiary-care referral center, which may reduce the generalizability of the results. Second, not all pre-1995 patients had right heart cardiac catheterization (RHC)-proven PH available for review. However, in all these cases the diagnosis and WHO classification had been confirmed by a UPMC Pulmonary Hypertension Center physician and all patients hadan estimated PASP >45mmHg at that time. Third, it is possible that patients with mild PH and PH secondary to ILD may have been missed in the first 15 years of patients included in this study, as both high-resolution computed tomography (CT) and echocardiograms were not routinely available for screening during that time. However, these omissions should have occurred at similar rates in the case and control populations given temporal matching of cases to the next two consecutive SSc patients seen in clinic. Finally, as we did not have computed tomography (CT) results or imaging available dating back to 1980 to allow quantification, we relied on the Medsger severity scale for ILD rather than formal CT quantification scores.

CONCLUSIONS

In long-term follow-up, anti-Th/To positive SSc patients have extremely high rates of developing PH, with an overall 37% frequency, of which WHO Group 1 PAH occurred most frequently (23%). Patients with anti-Th/To antibody typically present with limited or no skin thickening and a nucleolar pattern ANA. Such patients should undergo testing for anti-Th/To antibody and be vigilantly monitored for the development of PH and ILD.