Low Quality of Reports on Blood Pressure in Patients Adrenalectomized for Unilateral Primary Aldosteronism

Abstract

Context

Adrenalectomy is the preferred treatment for unilateral primary aldosteronism but the results of long-term control of blood pressure (BP) are far from optimal. One possible explanation relates to the quality of the assessment of treatment effects on BP.

Purpose of the study

To examine the quality of reporting BP measurements in studies assessing the outcome of adrenalectomy on BP.

Methods

We conducted a systematic review searching 3 databases (PubMed, EMBASE, Web of Science) for articles published from January 1, 1990, onwards. Sixty-six studies, each reporting on more than 50 adrenalectomized patients, were eligible for full analysis.

Results

In 37 of the analyzed 66 studies (56.1%) BP values both before and after adrenalectomy were reported. In 19.7% (13/66) of the studies the method of BP measurement was described. The number of visits and number of BP recordings per visit on which BP results were based were reported in <15% of papers. The criteria for the diagnosis of hypertension were described in 72.7% (48/66) of the studies. The used definitions of improvement of BP control after adrenalectomy were variable, with 84.8% of the studies not providing any quantitative criteria to define reduction in BP.

Conclusion

We conclude that the quality of reporting on BP control after adrenalectomy for primary aldosteronism shows substantial deficiencies and inconsistencies, thus impacting negatively on accurate assessment of effects of adrenalectomy on BP control. Future studies should adhere to accepted recommendations of correct BP measurement and should provide detailed description of the methods used for BP measurement.

aldosteronism, blood pressure, measurement, quality, outcome

Primary aldosteronism (PA) is an important cause of endocrine hypertension, affecting 1% to 29% of the hypertensive population (1). In most PA patients the adrenal cause of the elevated aldosterone secretion is a unilateral adenoma or bilateral hyperplasia (2). Patients with a unilateral adenoma are usually treated by adrenalectomy (ADX) while patients with bilateral hyperplasia are treated by mineralocorticoid receptor antagonists.

Adrenalectomy for an aldosterone-producing adenoma is very successful as aldosterone secretion is normalized in 90% to 100% of all operated patients (3,4). Since blood pressure (BP) is one of the most powerful determinants of long-term cardiovascular morbidity and mortality, surrogate clinical outcome is defined by the effects of ADX on BP, tied in with the amount of antihypertensive medication (5). In contrast to the excellent biochemical outcome, clinical results in terms of improved postoperative BP regulation is consistently reported to be far from optimal (3,6-8). A recent large multicenter study showed that in only 37% of the patients BP had completely normalized without antihypertensive medication while documenting a wide variation of 17% to 62% across centers (4). In 47% of the patients there was improvement of BP control.

Several potential explanations for these disappointing outcome results have been proposed. First, longstanding hypersecretion of aldosterone may have caused irreversible intra- and extrarenal vascular changes, thus impeding normalization of BP without antihypertensive medication in nearly 50% to 63% of the patients (3,4). Another explanation may be that some patients also suffer from underlying pre-existent primary hypertension as this is highly prevalent in the population. Finally, one might question whether the quality of assessment of BP control complies with the recommendations of the international guidelines on hypertension (9-11). Poorly performed measurement procedures of BP may obscure or confound the real effects on postoperative BP control. In addition, deficient and inconsistent BP measurement procedures across studies limit a reliable comparison of the effects of ADX on BP between studies. Finally, optimization of diagnostic procedures aims at improving outcomes for patients. It is therefore mandatory that BP outcome is properly assessed and reported in clinical studies.

In this systematic review we examined the protocols for and the reported measurement procedures that were used to assess the effects of surgery on postoperative BP in studies on patients who underwent ADX for unilateral PA. In addition, we examined the criteria that were applied to define normal BP and hypertension in studies on patients with PA before and after ADX.

Materials and Methods

Data sources and searches

The methods of this systematic review, including databases, search terms, and inclusion criteria were defined before the start of the review and were not modified thereafter. Three electronic databases were searched: PubMed, EMBASE, and Web of Science. We searched for articles in English, German, and French published from January 1, 1990, onwards. In addition, the reference lists of four previously published review papers were checked for additional relevant studies (3,6-8).

We used the following search terms: (“Hyperaldosteronism” [Mesh:noexp] OR Conn Syndrome [tiab] OR Conns Syndrome [tiab] OR Conn’s Syndrome [tiab] OR Primary Aldosteronism [tiab] OR Primary Hyperaldosteronism [tiab] OR aldosterone producing adenoma [tiab]) AND (“Adrenalectomy” [Mesh] OR Adrenalectomy [tiab] OR Adrenalectomies [tiab] OR Resection [tiab] OR Surgery [tiab]) AND (“Hypertension” [Mesh:noexp] OR “Prehypertension” [Mesh] OR “Blood Pressure” [Mesh:noexp] OR “Arterial Pressure” [Mesh] OR “Blood Pressure Determination” [Mesh:noexp] OR “Blood Pressure Monitoring, Ambulatory” [Mesh] OR Arterial pressure [tiab] OR Blood pressure[tiab] OR Diastolic Pressure[tiab] OR Hypertension [tiab] OR Normotension[tiab] OR Systolic pressure[tiab]). We merged search results from the 3 databases and checked automatically and manually for duplicates (J.L. and A.P.).

The literature search was last updated on June 20, 2019.

Study selection

Two researchers (J.L. and A.P.) independently assessed eligibility of retrieved articles based on title and abstract. If necessary, full-text articles were retrieved.

Studies were considered eligible for inclusion if they met the following criteria:

1. Original studies that had reported on outcome of BP after ADX in patients with unilateral PA. Reporting on BP outcome was not necessarily the predominant aim of a study.

2. Follow-up data on BP after ADX were available for the majority of patients.

3. Studies that reported on series of 50 or more patients who underwent unilateral ADX.

While several published studies included patients from different centers within a country, 6 of them were international multicenter studies. Some centers have published multiple studies, addressing different research questions over the time period covered by this review. Several centers had included, at least partly, the same patients in multiple studies as might be inferred from the reported time periods of patient inclusion. As we assumed that the procedures for assessment of BP had been the same for different studies from the same center(s), we chose the study with the largest number of evaluable patients if there was more than 1 study from a specific center in order to circumvent inclusion bias.

Disagreements on eligibility were resolved by consensus among the 2 reviewers (J.L. and A.P.).

Data extraction and quality assessment

Two researchers (J.L. and A.P.) independently scored all included studies and retrieved the following data: first author, year of publication, country, number of included patients who underwent ADX and in whom follow-up data were available, the time period over which patients were included in a study, reported BP values before and after ADX, time interval between ADX and BP assessment at follow-up, site of BP measurement (office, out-of-office at home, or 24-hour ambulatory BP measurement), method of BP measurement before and after ADX (sphygmomanometer or (semi-)automated device), body position in which BP was measured, duration of rest before measurement of office BP, number of BP measurements per visit, whether the first BP value for calculating the average BP per visit was discarded, the number of visits on which the reported pre- and postsurgical BP was based, and the criteria on which the diagnosis of hypertension was based. Reported cut-off levels of hypertension as ≥140/90 were interpreted as ≥140 and/or ≥90 mmHg unless stated otherwise. Finally, the criteria used for defining a reduction in BP were retrieved. Differences in extraction were resolved by consensus or, if necessary, by a third researcher (J.D.). Apart from the main text of the papers, we also analyzed the supplemental files when available.

Results

Results of the literature search are shown in the flow diagram (Fig. 1). We identified 1379 articles. Based on the exclusion criteria applied to titles and abstracts we performed a full-text analysis of the remaining 190 articles. An additional 53 articles had to be excluded for several reasons: articles reporting case reports or on fewer than 50 evaluable patients, review/commentary articles, other topics than follow-up after adrenalectomy. We excluded 71 additional papers because these papers were published by the same centers or groups and probably included at least partly the same patients as could be inferred from the reported inclusion periods. Finally, 66 studies, including a total of 9792 patients, were subjected to full qualitative and quantitative analysis (all supplementary material and figures are located in a digital research materials repository (12)).

Figure 1.

Flow diagram showing the steps of selection of studies, resulting in 66 studies eligible for full text analysis.

Reported BP values before and after ADX

In 35 of the 66 studies the systolic and diastolic BP values before and after ADX were reported while 2 studies reported mean arterial pressure. Fourteen studies reported only the presurgical baseline BP. Fifteen studies did not report BP values before or after ADX. In 56.1% of all studies BP values were reported both before and after ADX (Table 1).

Table 1.

Number of studies that reported specific items of BP measurement of the total cohort and of the studies published in the 3 decades

	n/N (%)	n/N (%)	n/N (%)	n/N (%)
Reported items of BP measurement	Total cohort	1990-1999	2000-2009	2010-2019
BP values both before and after ADX	37/66 (56.1)	2/7 (28.6)	4/8 (50.0)	31/51(60.8)
Site of BP measurementa	17/66 (25.8)	0/7 (0)	3/8 (37.5)	14/51 (27.5)
Method of BP measurementb	13/66 (19.7)	0/7 (0)	2/8 (25.0)	11/51 (21.6)
Position of the patient in which BP was measured	13/66 (19.7)	0/7 (0)	2/8 (25.0)	11/51 (21.6)
Time of rest before BP measurement	6/66 (9.1)	0/7 (0)	1/8 (12.5)	5/51 (9.8)
Number of BP measurements per visit	10/66 (14.9)	0/7 (0)	2/8 (25.0)	8/51 (15.7)
Discarding first BP value for calculating average BP per visit	3/66 (4.5)	0/7 (0)	1/8 (12.5)	2/51 (3.9)
Number of visits on which the presurgical BP was basedc	2/51 (3.9)	0/4 (0)	1/4 (25)	1/43 (2.3)
Number of visits on which the postsurgical BP was basedd	4/37 (10.8)	0/4 (0)	2/4 (50.0)	2/31 (6.5)
Time interval between ADX and BP measurement at follow-up	61/66 (92.4)	6/7 (85.7)	8/8 (100)	47/51 (92.1)

BP, blood pressure; ADX, adrenalectomy; n, number of studies; N, total number of studies.

^aOffice or out-of-office (home or 24-hour ambulatory BP measurement).

^bSphygmomanometer or (semi-)automated device.

^cIncluding the patients in whom only the presurgical BP values were reported.

^dFor the patients in whom both pre- and postsurgical blood pressure values were reported.

Procedure of BP measurement

Seventeen of the 66 studies (25.8%) reported that BP was measured in the office, including 2 studies reporting that additional 24-hour ambulatory BP monitoring was used. In 1 study, home BP measurement was used in addition to office BP. A majority of 49 studies (74.2%) did not report the site of BP measurement (Table 1).

Eight studies mentioned the use of a sphygmomanometer, including 5 studies in which use of a (semi-)automatic electronic device was also reported but it was not clarified which method was used for the assessment of BP outcome. Five studies stated the use of a (semi-)automatic electronic device. Of all 66 studies, the majority of 53 (80.3%) did not report the method of BP measurement (Table 1).

Five studies reported a resting time of 5 minutes preceding BP measurement, while 1 study employed 10 minutes of rest. Sixty studies (90.9%) did not clarify whether any specific resting time was installed before BP measurement (Table 1).

According to guideline recommendations, the body position of the patient in which the BP is measured is also relevant. Thirteen studies described the position of the patients when BP was measured. Nine studies followed the guideline recommendation of measuring BP in the sitting position while in 4 studies it was measured with the patient supine. In 53 studies (80.3%) no information about the position of the patient was provided (Table 1).

Schedule of BP measurements

The number of visits on which reported BP was based was described in a limited number of studies: 1 study stated that the presurgical BP was based on 3 visits while in 1 study this was based on only 1 visit. Postsurgical BP appeared to be based on 1 visit in 5 studies while this was based on 2 visits in another 2 studies. In the majority of studies in which the pre- and postsurgical BP data were provided, the number of visits on which the reported presurgical (35/51) and postsurgical (35/37) BP values were based was not reported (Table 1).

Ten studies reported the recording of 1 or more BP measurements during a visit. One study reported 10 BP recordings, 6 studies employed 3 BP recordings, 1 study 4, 1 study more than 2, and 1 study reported only 1 recording. Of these 10 studies, 3 described that the first BP reading was discarded for calculating the average BP of a visit. One study used the only 1 BP value that was recorded. In 56 studies (84.8%) we could not identify a description of how many BP recordings were carried out per visit and how the average BP of that visit was calculated (Table 1).

Some studies described follow-up evaluation at regular time intervals after ADX, but we have used only the time interval that was used for the assessment of outcome. Most studies reported the time interval between ADX and follow-up visit to assess BP (Table 1). In 25 studies the time period that elapsed after ADX till follow-up assessment of BP was reported as a certain predefined time interval, varying from 0.25 to 168 months. In 18 studies the reported actual mean time intervals varied from 10.8 to 106 months across studies while in 18 studies the median time intervals (varying from 0.66 to 144 months) were provided. Five studies did not report the time interval of follow-up.

Analysis of all data over time showed that the proportion of studies reporting on all items concerning BP measurement was lowest in those published between 1990 and 1999 while in the last 2 decades of 2000 to 2009 and 2010 to 2019, there was clear improvement for 8 of all 10 items that were evaluated (Table 1). Nevertheless, in the last decade from 2010 to 2019, the proportion of studies reporting 8 of the 10 evaluated items remained less than 30%.

Criteria used to define hypertension and BP improvement

In most studies the definition of cure of hypertension after ADX was reported and we assumed that for presurgical BP assessment the same definition of hypertension was used. Several studies (18/66) did not provide explicit criteria to define the diagnosis (pre- or postsurgical) hypertension while a few studies referred to international guidelines. In 48 studies the following variety of criteria for systolic and diastolic BP respectively were used: ≥140 and/or ≥90 (n = 30); >140 and/or >90 (n = 6), ≥140 and ≥90 (n = 3); >140 and >90 (n = 1); >160 or >100 (n = 1); >160/95 (n = 2); some used only systolic BP >140 (n = 2) or >130 (n = 1), diastolic BP >100 (n = 1) or MAP (mean arterial blood pressure) >110 (n = 1).

Regarding assessment of outcome of BP after ADX, 7 studies applied the recently described PASO (Primary Aldosteronism Surgery Outcome) criteria (4). Thirty-six studies defined cure of hypertension as a BP of <140/90 with no use of antihypertensive drugs. A few studies also used the term “normotension” without specifying the criteria of normotension. The definition of improvement of BP control was considerably variable with most studies (84.8%) not providing any preset quantitative criteria of BP reduction. Most reports just described improved BP control as “similar or lower BP compared with presurgical BP with either use of fewer or a similar number of antihypertensives” without any quantitative specification.

Reporting on medical treatment

Although most studies reported in their Methods sections that they recorded antihypertensive treatment, in only 30 of the 37 studies reporting on the BP effects before and after adrenalectomy were the numbers of used antihypertensive drugs provided. In 20 of these studies the reduction in number of antihypertensive drugs was also reported. Two of the 37 studies also provided the intensity of drug dosing (as daily defined doses).

Discussion

Our study shows dismal assessment and reporting of BP outcomes in studies of patients who underwent ADX for unilateral aldosterone-producing adenoma. The majority of studies did not provide methodological details of BP measurement and quantitative criteria to estimate the effects of ADX on BP control. Inconsistent use and reporting of the definition of hypertension is another shortcoming. Although the results tended to improve over time, the majority of most recently published studies did not provide sufficient detail regarding the methodology of BP measurement.

In particular, only a small minority reported the device used for measuring BP. In addition, most studies did not describe other procedural aspects of measurement of BP such as site of measurement, position of the patient during measurement, preceding resting interval before measurement, number of visits, and number of BP recordings per visit on which the given BP results are based. Across studies, the time interval between ADX and follow-up varied considerably from a few weeks to several years. We have assumed that the method used for BP measurement was the same before and after ADX, although only 5 studies stated this explicitly. Finally, descriptions of improvement of BP control were variable but most studies did not provide any quantitative criteria to define reduction in BP.

Methodological aspects of BP measurement are critical for a reliable estimation of the real BP, not only for clinical care but also for clinical research studies in which BP is used as the outcome parameter. This is the reason that several international guidelines have published recommendations on how BP and effects of interventions on BP should be measured (9-11). A recently published report from a scientific expert panel of the National Heart, Lung, and Blood Institute emphasized the importance of using rigorous standardized protocols for BP measurement to ensure comparability and accuracy of BP in clinical studies (11). A minimum standard for conducting and reporting BP measurements in clinical studies has been proposed recently (13). In Table 2 we have summarized the most important recommendations for BP measurement in the context of clinical studies. In our study we have addressed not all but several of these most important requirements for proper BP measurement as recommended by the guidelines, as we expected that some items were not addressed at all, such as calibration and validation of devices and the appropriate position and size of the upper arm cuff.

Table 2.

Basic requirements for reporting BP levels in interventional clinical studies in which blood pressure control is one of the main outcome parameters

Minimum standard for conducting and reporting BP in clinical studiesa	Evaluated in the current study
Trained personnel	No
Available BP measurement protocol	No
Site of BP measurement: office, ambulatory (home or 24-hour ABPM)	Yes
Position of patient during measurement: seated, supine, standing	Yes
Resting period before measurement	Yes
Description of measurement method: sphygmomanometer; (semi-)automated device	Yes
Independent validation and calibration of used devices	No
Appropriately positioned upper arm cuff of adequate size/arm supported at heart level	No
Seated position with the back of the patient supported and legs uncrossed	No
Multiple readings per visit and multiple visits at baseline and during postinterventional follow-up	Yes
Calculation of average BP per visit and discarding the first BP value	Yes
Follow-up time interval between intervention and BP assessment	Yes

BP, blood pressure; ABPM, ambulatory blood pressure measurement.

aMuntner et al. J Am Coll Cardiol. 2019;73:317-335 11 and True Consortium. Clin Exp Hypertens. 2018;40:509-513 13.

Relying on BP results from only electronic medical records without a standardized protocol for BP measurement will introduce systematic and random errors. Systolic BP data from clinical research studies employing a standardized protocol have been shown to be on average 6.5 mm Hg (95% CI, 4.2-7.8) lower than the most proximal clinic BP derived from the medical records (14). A recent study also demonstrated that the crucial relevance of trial design in interventional studies as nonrandomized unblinded studies showed large reductions in office BP compared with no reduction in BP when using ambulatory BP measurements (15).

These data argue for ambulatory BP measurement for drug trials, but it seems reasonable to extend this to other interventions such as ADX. Different methods of BP measurement impact considerably the accurate assessment of BP, for example, different positions in which BP is measured, no or different periods of preceding rest, number of BP recordings per visit, number of visits, whether or not the first measurement for calculating the mean BP of a visit is discarded. It is therefore disappointing that our study clearly shows that these recommendations are ignored by most studies that deal with BP as one of the relevant outcome parameters of ADX. Accurate, reliable, and standardized measurement of BP is a prerequisite to obtain a real and accurate estimation of the treatment effects of ADX. This is of basic importance as BP is a crucial determinant of long-term cardiovascular and renal outcome (16).

In addition to the above-listed shortcomings in the reported measurement procedure itself, there are also deficiencies in the criteria used to define hypertension. Nearly 1 in 4 studies did not describe any criterion defining pre- or postsurgical hypertension. The variable criteria in the other studies illustrate that the diagnostic criteria for hypertension were not applied in a uniform way. Consequently, it is difficult to estimate the real effects of ADX on BP, in particular the proportion cured of hypertension.

Except for cure of hypertension, improvement of BP control after ADX (lower BP and/or lower intensity of antihypertensive medication and none of them higher) will also have a beneficial impact on the prognosis of patients. Our study indicates not only a large variability of criteria to define the effect on BP after ADX but also most studies did not report whether and how this effect was quantified. The need for more standardized criteria on how to define the effects of ADX on BP is met by the recently published PASO study, which has provided a consensus on standardized and quantitative criteria for assessing outcome after ADX (4). Stringent adherence to these criteria will enable better comparison between studies and therefore provide a more reliable estimate of BP improvement in those patients in whom hypertension is not cured.

Our study has some limitations. The collected data on how BP was assessed are based on what has been reported in the published papers. This is not necessarily identical to how BP was really measured in the study. In addition, it may well be that those studies that did not report on the measurement procedure in reality adhered better to the accepted recommendations than it seems. Despite these caveats, the result is that the reported data on outcome are not comparable between studies. Some studies described procedures in supplementary files and these have been included in the analysis.

Recommendations on definition of hypertension and on use of out-of-office BP measurements might have evolved over the wide time span that is covered by our study. However, in the early 1990s hypertension was defined as ≥140 and/or ≥90 mmHg and only recently has the US guideline downgraded the definition of elevated BP to a systolic BP between 120 and 129 and diastolic less than 80 mmHg (9,10). In the early nineties, out-of-office measurements such as home and 24-hour ambulatory BP measurements were recommended only to be used in selected cases to complement office BP measurement. Nowadays, out-of-office BP measurements are recommended to confirm the diagnosis of hypertension (10). The problem is however not the lack of out-of-office BP measurements when evaluating previous studies but the lack of proper description at which sites BP has been measured.

Finally, a few studies referred to guidelines without specifying what was actually done in their study patients.

We conclude that the quality of reporting on outcome of BP control after ADX for unilateral primary aldosteronism shows several serious deficiencies and inconsistencies. This is concerning as it not only impacts negatively on accurate assessment of the effects of adrenalectomy on BP control but also renders a meaningful comparison of BP outcome across studies hardly possible. Future studies should provide a detailed description of the methods used for BP measurement and adhere better to the accepted recommendations of correct BP measurement.

Abbreviations

Abbreviations
 
• ADX

  adrenalectomy

 
• BP

  blood pressure

 
• PA

  primary aldosteronism

Acknowledgments

Financial Support: J.W.M.L. and J.D. were supported by the Deutsche Forschungsgemeinschaft (DFG) within the CTC/Transregio 205/1, project B15, “The Adrenal: Central Relay in Health and Disease.

Additional Information

Disclosure Summary: The authors have nothing to disclose. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Data Availability: The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.